Proceedings of the 5th Nordic Feed
Science Conference, Uppsala, Sweden
Institutionen för husdjurens
utfodring och vård
Rapport 290
Report
Swedish University of Agricultural Sciences
Department of Animal Nutrition and Management
Uppsala 2014
ISSN 0347-9838
ISRN SLU-HUV-R-290-SE
Proceedings of the 5th Nordic Feed
Science Conference, Uppsala, Sweden
Institutionen för husdjurens
utfodring och vård
Rapport 290
Report
Swedish University of Agricultural Sciences
Department of Animal Nutrition and Management
Uppsala 2014
ISSN 0347-9838
ISRN SLU-HUV-R-290-SE
Published by:
Organising committee of the 5th Nordic Feed Science Conference
Department of Animal Nutrition and Management
Swedish University of Agricultural Sciences (SLU)
SE- 753 23 Uppsala, Sweden
Copyright © 2014 SLU
All rights reserved. Nothing from this publication may be reproduced, stored in computerised
systems or published in any form or any manner, including electronic, mechanical, reprographic or
photographic, without prior written permission from the publisher (SLU).
The individual contributions in this publication and any liabilities arising from them remain the
responsibility of the authors.
Organising Committee:
Peter Udén
Rolf Spörndly
Torsten Eriksson
Cecilia E. Müller
Marie Liljeholm
Editor in chief:
P. Udén
Editorial board:
T. Eriksson
B-O. Rustas
C.E. Müller
R. Spörndly
T. Pauly
M. Emanuelson
K. Holtenius
M. Patel
M. Åkerlind
A. Jansson
Printed by:
SLU Repro
SE-75007 Uppsala, Sweden
Distributed by:
Department of Animal Nutrition and Management,
Kungsängen Research Centre
SE-75323 Uppsala, Sweden
www.slu.se/husdjur-utfodring-vard
QR code reader for accessing conference posters can be downloaded at:
http://app.scanlife.com/appdownload/dl
Proceedings from the Nordic Feed Science conferences are indexed and archived in CAB Abstracts
since 2010
Foreword
This year we celebrate the 5th anniversary of the Nordic Feed Science Conference which
started in 2010. The aim of the conference has been to create an arena for Nordic feed
scientists to meet and discuss ruminant and horse feeds and feeding. The 1st year, the
conference drew a high number of participants (119) and after that, it has stabilized at about
60 (Figure 1). The number of contributions has varied from 25 to 44 and, outside Sweden,
Denmark and Finland have been the most frequent contributors (Table 1). It is gratifying that
in the last years, we have seen 2 to 3 joint Nordic papers. Maybe, this is a result of more
cooperation among the Nordic countries.
120
100
80
60
40
20
0
Participants
2010
119
2011
59
2012
59
2013
45
2014
60
Figure 1 Number of participants
Table 1. Contribution to the Conference from different countries
Year
SE
DK
FI
NO
ISL
NORD
EST
LV
LT
Other
Total
2010
21
7
7
3
0
0
1
0
0
5
44
2011
17
4
2
0
0
0
0
2
0
0
25
2012
14
0
1
4
1
0
0
0
0
11
31
2013
15
4
1
1
0
2
1
0
0
0
24
2014
12
2
8
0
0
3
1
0
2
5
33
Mean
15.8
3.4
3.8
1.6
0.2
1.0
0.6
0.4
0.4
4.2
31.4
One topic of particular interest this year is protein evaluation. In ruminants, Session II deals
exclusively with protein nutrition and in the evening session, we anticipate interesting
discussions of the value of forage proteins. The two contributions on horses also deal with
protein.
This year we are extremely happy to have three internationally renowned scientists speaking
to us. Limin Kung from the Dairy Nutrition and Silage Fermentation Lab, University of
Delaware, Karl-Heinz Südekum from the University of Bonn and (for the 4th time) Glen
Broderick from Madison, Wisconsin. We wish to welcome these eminent scientists for
coming all the way to join us in this conference.
Uppsala 2014-06-02
Peter Udén
Proceedings of the 5th Nordic Feed Science Conference
Contents
Forage conservation and feed processing
Ecology of silage microorganisms
K. Mogodiniyai Kasmaei, V. Passoth, P. Udén
5
Managing the aerobic stability of silages
L. Kung Jr.
10
Aerobic exposure of silages – effects on chemical composition and
preference by goats
K.-H. Südekum, K. Gerlach
15
Effects of particle size and chemical additives on fermentation and
aerobic stability of grass-clover silage
E. Nadeau, H. Auerbach
19
Effects of different moisture contents and additives on the quality
of baled oat silage
Z. Gui-Qin, Q. Fang-Cuo, J. Ting, H. Jian-Jie, S. Xu-Dong
25
The effect of silage additive on fermentation quality, clostridia and
aerobic stability of a white lupin-wheat silage
W. König, L. Puhakka, M. Lamminen, K. Weiss, A. Vanhatalo, S. Jaakkola
31
Protein value in organic grass-clover silages from spring and summer growth
A.K. Bakken, M. Vaga, M. Hetta, Å. Taksdal Randby, H. Steinshamn
37
Dynamics of gasformation during ensilage
M. Knicky, H-G. Wiberg, F. Eide, B. Gertzell
41
Mycotoxins in haylage
J. Schenk, C. Müller, R. Spörndly
47
Energy efficient and competitive on farm storage of moist feed grain –
preliminary results
N. Jonsson, J. Blomqvist, M. Olstorpe
Agro-food industry wastes conversion into safer feed stock by using
lactic acid bacteria
V. Krungleviciute, E. Bartkiene, J. Kantautaite, G. Juodeikiene, J. Damasius,
V. Baliukoniene, B. Bakutis
51
57
Ruminant nutrition
Should rumen microbial protein formation be maximized?
G. A. Broderick
58
Biological limits to N utilization in dairy cows
S. Ahvenjärvi, T. Stefanski, A. Vanhatalo, P. Huhtanen
67
Evaluation of protein supplementation for growing cattle: A meta-analysis
A. Huuskonen, P. Huhtanen, E. Joki-Tokola
72
Proceedings of the 5th Nordic Feed Science Conference
A meta-analysis of milk production responses to increased supply of AAT
I. Schei, N. I. Nielsen, M. Åkerlind, Volden, H.
78
Dietary nutrient density and effects on intake and production
S. Rengman, B. Johansson, M. Murphy
83
Intra-ruminal mixing of concentrate pellets produced by conventional pellet
press or extruder
M. Nordqvist, P. Lund, A. C. Storm, M. R. Weisbjerg, M. Larsen
88
Prediction of methane emissions from dairy cows fed cold-pressed linseed cake
M. Kass, A. Olt, R. Leming, M. Ots
92
Effects of concentrate feeding strategies on performance of dairy bulls
K. Manni, M. Rinne, A. Huuskonen
97
Effects of supplementary concentrate level and separate or total mixed ration
feeding on performance of growing dairy bulls
M. Pesonen, E. Joki-Tokola, A. Huuskonen
103
Effect of salt addition to a barley concentrate on milking frequency,
milk yield and feed intake in automatic milking systems
M. Johansen, M. Larsen, P. Lund & M. R. Weisbjerg
109
The evaluation of efficiency of different mineral additives on milk yield and
quality in dairy cows
V. Jokubauskienė, V. Špakauskas
113
Selenium supplementation by addition of sodium selenate with silage additive
A. Seppälä, Y. Madrid Albarran, H. Miettinen, M. Palomo Siguero,
E. Juutinen, M. Rinne
118
Factors influencing the production value of forage protein
P. Huhtanen
124
Horses
The new protein evaluation of horse feeds in Germany
K-H Südekum
130
Effects of crude protein content in forage-only diets fed to horses
A. Jansson, S. Ringmark
133
Methods and miscellaneous
Simple Excel VBA application for Monte Carlo simulation with
ad hoc made spreadsheet models
T. Eriksson
138
Comparison between individual feeds and total diets on predicted
methane production from in vitro simulations
M. Ramin, M. Vaga, E. H. Cabezas-Garcia and E. Detmann
143
Proceedings of the 5th Nordic Feed Science Conference
Relationship between chewing index and intake of metabolizable energy
in pregnant ewes
M. Vestergaard Nielsen, E. Nadeau, B. Markussen, C. Helander,
M. Eknæs, Å. Randby, P. Nørgaard
In-line oxygen monitoring in lab-scale silos
T. Pauly
Effects of heat treatment on protein feeds evaluated in vitro utilising
the method utilisable protein
M. Vaga, M. Hetta, P. Huhtanen
Grass silage extract, feed component suitable for pigs –
prospects for on farm biorefinery
A. Seppälä, S. Kyntäjä, L. Blasco, M. Siika-Aho, S. Hautala, O. Byman,
H. Ilvesniemi, H. Ojamo, M. Rinne, M. Harju
Dry matter yields and feed values of faba bean and pea as bi-crops with
wheat or oats in Northern Finland
K. Kuoppala, A. Huuskonen, E. Saarinen, M. Rinne
147
153
157
163
169
Proceedings of the 5th Nordic Feed Science Conference
Forage conservation and feed processing
Ecology of silage microorganisms
K. Mogodiniyai Kasmaei1, V. Passoth2 & Peter Udén1
1
Swedish University of Agricultural Sciences (SLU), Department of Animal Nutrition &
Management, Feed Science Division, Kungsängen Research Centre, 753 23 Uppsala,
Sweden. 2 Swedish University of Agricultural Sciences (SLU), Department of Microbiology,
756 51 Uppsala, Sweden
Correspondence: [email protected]
Introduction
The biochemistry and microbiology of ensiling have been investigated in depth in several
studies (McDonald et al., 1991; Rooke and Hatfield, 2003; Pahlow et al., 2003). It is,
however, difficult to combine these results to address as to why the catabolic diversity in
silage ecosystem is large and what are the driving forces for such complexity. As an
example, certain lactic acid bacteria under the presence of air reduce O2, by which, reactive
oxygen species (e.g. superoxide, hydrogen peroxide) which are essentially detrimental to
them are formed (Pahlow, 1991). In this work, we aim at dissecting factors of importance to
the silage ecosystem and provide an overview of some of the catabolic pathways employed
by silage microorganisms.
Respiration versus fermentation
Understanding similarities and differences between respiration and fermentation is essential
for understanding causes and effects in the silage ecosystem. In both processes, chemical
energy of substrates is released by means of oxidation-reduction (redox) reactions. Electrons
are transferred from substrates to compounds with higher reduction potentials, i.e. the
electron acceptors, by electron carrier molecules (e.g. nicotinamide adenine dinucleotide). A
larger difference between reduction potential of electron donors and acceptors means a
greater magnitude of the ΔG0'. In biological systems, O2 has the highest reduction potential
(Madigan et al., 2012). The half-reactions, i.e. oxidation of the substrate and reduction of the
electron acceptor, always take place together so that the electron carrier molecules become
again available and the process of energy release continues, a phenomenon referred as a
balanced reaction (Madigan et al., 2012). It is therefore appeared that this process is heavily
dependent on the presence of electron acceptor compounds in the system.
In fermentative pathways, the redox balance is reached by reduction of intermediates (e.g.
pyruvate) that are formed from the up-taken substrates. This results in an incomplete
oxidation and excretion of semi-catabolized substrates, i.e. fermentation end-products. In
respiratory pathways, external electron acceptors (EEAs), e.g. O2, or nitrate, are exploited
instead. A further distinction between the two is that in fermentation, the released energy is
stored via substrate level phosphorylation but in respiration, oxidative phosphorylation is also
employed. This in turn gives substantial superiority to respiration regarding energy
conservation efficiency.
Silage ecosystem
A low pH of silage causes stress to acid tolerant microorganisms while being detrimental to
the others. This effect is the result of passive entry of short chain fatty acids (e.g. acetic and
propionic acids) into the cell before dissociation and decreasing the intracellular pH to
detrimental levels. Silage microorganisms employ different strategies to withstand a low pH.
Active removal of H+ is the most important defensive response (De Angelis and Gobbetti,
Proceedings of the 5th Nordic Feed Science Conference
5
Forage conservation and feed processing
2004). It seems, therefore, under low pH conditions, silage microorganisms are in a greater
demand of energy.
When air penetrates into a silo, growth of yeasts is enhanced (Jonsson and Pahlow, 1984).
This increased population growth has been attributed to the lactate assimilating ability of
certain species, collectively known as lactate assimilating yeasts. Firstly, such classification
of silage yeasts (i.e. lactate vs. non-lactate assimilating) is an approximation because the
response of different yeast species to lactate depends heavily on environmental conditions
(e.g. N source and pH) and adaptation time (Middelhoven and Franzen, 1986). Secondly,
other fermentation end-products, such as ethanol, can also be assimilated by yeasts when air
is present. Therefore, this classification is incomplete and could be misleading. The increased
growth is, however, explained by the fact that respiratory species (e.g. Wickerhamomyces
anomalus) can switch from fermentation to respiration (Madigan et al., 2012), by which, the
efficiency of energy conservation is improved.
A similar strategy of increasing energy gain can also be found in prokaryotic species. For
instance, Escherichia coli is able to anaerobically respire when nitrate is available (Madigan
et al., 2012). Leuconostoc spp. and Lactobacillus plantarum will ferment glucose to acetic
acid instead of lactic acid under the presence of O2 and nitrate, respectively (Rooke and
Hatfield, 2003). Through this externally balanced fermentative pathway, two extra moles of
ATP are gained. Lactic acid bacteria generally lack catalase for detoxification of reactive
oxygen species (Pahlow, 1991; Sanders et al., 1999) but as discussed in the Introduction, they
still opt reducing O2. This phenomenon might be explained by the theory that silage
microorganisms prioritize enhancement of energy gain due to an inflated energy demand. On
standing forges, this inflation is owing to starvation, irradiation and climatic conditions and
during ensiling, it is caused by the low pH.
Anaerobic condition together with the closed system of ensiling favor lactic acid bacteria as
these conditions allow accumulation of lactic acid and, thereby, out-competition of pH
sensitive microorganisms. The ability of silage microorganisms to produce anti-microbial
compounds that directly target competitors, however, should not be ruled out. For instance, it
is known that L. plantarum, L. buchneri and W. Anomalus have the ability to produce certain
anti-microbial compounds (Gollop et al., 2005; Passoth et al., 2006; Olstorpe et al., 2011).
Different strategies used by silage microorganisms for increasing energy conservation
efficiency can be summarized as shown in Figure 1.
6
Proceedings of the 5th Nordic Feed Science Conference
Forage conservation and feed processing
Figure 1 A summary over catabolic pathways used by silage microorganisms.
The metabolic ability of microorganism (e.g. yeasts cannot practice anaerobic respiration)
and redox balance status (i.e. availability of EEAs) are two factors that determine pathway.
Substrate availability and silage pH can also play important roles (McDonald et al., 1991).
The complexity is even further increased by the fact that some microorganisms can utilize
certain end-products during starvation. It was, for instance, found that under the presence of
citrate, L. plantarum can ferment lactic acid, using citrate as the EEA (Lindgren et al., 1990).
Implications
Modeling silage quality from pre-ensiled forage composition could be of great benefit to
farmers. Advice on the correct kind of additives, wilting strategies, etc. could eventually be
improved. Such models can also play important parts in silage research and additive
evaluations. However, most of the attempts in this regard (Wilkinson et al., 1983; Pitt et al.,
1985; Mogodiniyai Kasmaei et al., 2013) have been largely unsuccessful. The majority of
these modeling attempts, including both dynamic and static models, have not taken the role of
microbial interactions into account, something which may have contributed to their failures.
This has mainly been due to scarcity of the relevant data and/or underestimating the effect of
this factor on ensiling results. For constructing more powerful silage models, data on
microbial interaction seem to be absolutely necessary.
Improving the aerobic stability of silage upon opening has recently become of great interest.
Inoculation with L. buchneri, a heterofermentative species, has been suggested (Kleinschmit
and Kung, 2006). However, this strategy could lead to an increased dry matter loss, which is
explained by a slow rate of pH decline that allows continuation of microbial catabolic
activities and formation of CO2 by heterolactic fermentation (Wilkinson and Davies, 2012).
Application of chemical additives can be an alternative strategy to improve both aerobic
stability and DM losses (Knicky and Spörndly, 2011). However, the cost and unsustainability
are unfortunately disadvantages of this technique (Wilkinson and Davies, 2012). An
increased knowledge of microbial ecology could improve silage research and result in a new
generation of additives or even new storage strategies.
Proceedings of the 5th Nordic Feed Science Conference
7
Forage conservation and feed processing
Conclusions
Silage microorganisms strive to enhance energy conservation efficiency. The strategy used is
a constant adjustment of catabolic pathways to environmental conditions including substrate
availability, redox balance and pH. A better understanding of microbial ecology is needed to
improve silage research and technology.
References
De Angelis, M. & Gobbetti, M., 2004. Environmental stress responses in Lactobacillus: A
review. Proteomics 4, 106-122.
Gollop, N., Zakin, V. & Weinberg, Z.G., 2005. Antibacterial activity of lactic acid bacteria
included in inoculants for silage and in silages treated with these inoculants. J. Appl.
Microbiol. 98, 662-666.
Jonsson, A. & Pahlow, G., 1984. Systematic classification and biochemical characterization
of yeasts growing in grass silage inoculated with Lactobacillus cultures. Animal Research
and Development 20, 7-22.
Kleinschmit, D.H. & Kung, L. Jr., 2006. A meta-analysis of the effects of Lactobacillus
buchneri on the fermentation and aerobic stability of corn and grass and small-grain silages.
J. Dairy Sci. 89, 4005-4013.
Knicky, M. & Spörndly, R., 2011. The ensiling capability of a mixture of sodium benzoate,
potassium sorbate, and sodium nitrite. J. Dairy Sci. 94, 824-831.
Lindgren, S.E., Axelsson, L.T. & McFeeters, R.F., 1990. Anaerobic L-Lactate Degradation
by Lactobacillus-Plantarum. FEMS Microbiol. Lett. 66, 209-213.
Madigan, M.T., Martinko, J.M., Stahl, D.A. & Clark, D.P., 2012. Brock Biology of
Microorganisms. Pearson Education, San Francisco.
McDonald, P., Henderson, A.R. & Heron, S.J.E., 1991. The Biochemistry of Silage.
Chalcombe publications, Marlow, Bucks, UK.
Middelhoven, W.J. & Franzen, M.M., 1986. The Yeast flora of ensiled whole-crop maize. J.
Sci. Food Agr. 37, 855-861.
Mogodiniyai Kasmaei, K., Rustas, B.-O., Spörndly, R. & Udén, P., 2013. Prediction models
of silage fermentation products on crop composition under strict anaerobic conditions: A
meta-analysis. J. Dairy Sci. 96, 6644-6649.
Olstorpe, M., Jacobsson, K., Passoth, V. & Schnürer, J., 2011. Controlling fungal growth and
mycotoxins in animal feed, In: Lacroix, C. (Ed.), Protective Cultures, Antimicrobial
Metabolites and Bacteriophages for Food and Beverage Biopreservation. Woodhead
Publishing Ltd., Cambridge, UK, pp. 225-239.
Pahlow, G., 1991. Role of microflora in forage conservation, In: Pahlow, G., Honig, H.
(Eds.), Forage conservation towards 2000. Landbauforschung Völkenrode, Bruanschweig,
Germany, pp. 26-36.
Pahlow, G., Muck, R.E., Driehuis, F., Oude Elferink, S.J.W.H. & Spoelstra, S.F., 2003.
Microbiology of ensiling, In: Buxton, D.R., Muck, R.E., Harrison, J.E. (Eds.), Silage Science
and Technology. Am. Soc.Agron.Inc., Madison, Wisconsin, USA, pp. 31-93.
8
Proceedings of the 5th Nordic Feed Science Conference
Forage conservation and feed processing
Passoth, V., Fredlund, E., Druvefors, U.A. & Schnürer, J., 2006. Biotechnology, physiology
and genetics of the yeast Pichia anomala. FEMS Yeast Res. 6, 3-13.
Pitt, R.E., Muck, R.E. & Leibensperger, R.Y., 1985. A quantitative model of the ensilage
process in lactate silages. Grass Forage Sci. 40, 279-303.
Rooke, J.A. & Hatfield, R.d., 2003. Biochemistry of ensiling, In: Buxton, D.R., Muck, R.E.,
Harrison, J.E. (Eds.), Silage Science and Technology. Am. Soc.Agron.Inc., Madison,
Wisconsin, USA, pp. 95-139.
Sanders, J.W., Venema, G. & Kok, J., 1999. Environmental stress responses in Lactococcus
lactis. FEMS Microbiol Rev. 23, 483-501.
Wilkinson, J.M., Chapman, P.F., Wilkins, R.J. & Wilson, R.F., 1983. Interrelationships
between pattern of fermentation during ensilage and initial crop composition, In: Smith, J.A.,
Hays, V.W. (Eds.), Proc.XIV Intern. Grassl. Congr., Kentucky, USA, pp. 631-634.
Wilkinson, J.M. & Davies, D.R., 2012. The aerobic stability of silage: key findings and
recent developments. Grass Forage Sci. 68, 1-19.
Proceedings of the 5th Nordic Feed Science Conference
9
Forage conservation and feed processing
Managing the aerobic stability of silages
L. Kung Jr.
Dairy Nutrition & Silage Fermentation Laboratory, Department of Animal & Food Sciences,
University of Delaware, Newark, Delaware 19716, USA
Correspondence: [email protected]
Introduction
Prolonged infiltration of air during storage or feed out into the silage mass can lead to aerobic
spoilage. Silage that is unstable when exposed to air heats rapidly and spoils, leading to a
loss of DM and nutrients with the potential for production of undesirable compounds.
Aerobic spoilage during storage often is responsible for a large portion of the total DM loss in
forages conserved as silage. Losses of DM could be as high as 15 to 25% or even higher in
areas of poor compaction. This review will focus on the processes involved during aerobic
spoilage of silages, why aerobic spoilage is undesirable and ways to improve aerobic
stability.
Aerobic stability of silages
When the active stage of ensiling is completed, the remaining microorganisms in the mass are
relatively dormant because of the low pH and absence of oxygen. However, if silage is
exposed to air, the result can be a chain reaction resulting in aerobic spoilage. Specifically,
yeasts that are able to degrade lactic acid in the presence of air usually initiate this process.
Examples of these organisms include Candida krusei (Issatchenkia orientalis) and Pichia
membranifaciens (C. valida) (Inglis et al., 1999). Yeasts able to metabolize sugars (e.g.
Saccharomyces) are also active and can exacerbate the spoilage process. Degradation of
lactic acid specifically causes an increase in pH of the silage to a level that allows
opportunistic bacteria (e.g. Bacilli) and molds (e.g. Aspergillus, Fusarium, and Pencillium) to
then become active, furthering the spoilage process (McDonald et al., 1991). Aerobic
microbial activity causes oxidation of nutrients resulting in the production of heat. Of
particular concern is the development of pathogenic bacteria and molds in aerobic spoiling
silages because of their potential for producing mycotoxins and causing other detrimental
effects. In some cases, bacteria from the genus, Acetobacter may initiate aerobic spoilage in
maize silages (Spoelstra et al., 1988). Grain crops (e.g. whole crop barley silages, maize
silages, and high moisture maize-grain silage) are very prone to aerobic spoilage. Epiphytic
populations of yeasts are found on all forage crops in the field. However, their numbers are
not well correlated to aerobic stability because there is a mix of non-fermentative and
fermentative species present. The ensuing fermentation and level of silage management
ultimately determine the number of lactate-assimilating yeasts that may survive ensiling. A
list of some factors affecting the aerobic stability of silages is shown in Table 1. The ambient
temperature around the silage mass affects the rate of aerobic spoilage. When temperature is
low, microbial activity is slowed or even stopped (e.g. in freezing weather). Warm climate
stimulates microbial activity and thus, aerobic spoilage and it is the primary reason that more
spoilage typically occurs in the summer than in the winter. High concentrations of residual
sugars in silage can also lead to a higher probability of aerobic spoilage.
Impact of feeding aerobically spoiled silages to ruminants
Few studies have been conducted evaluating the effects of feeding aerobically spoiled silages
to ruminants. Whitlock et al. (2000) reported that feeding spoiled corn silage from the
10
Proceedings of the 5th Nordic Feed Science Conference
Forage conservation and feed processing
surface of a bunker silo depressed DM intake as the level of spoiled feed in the diet increased
from 0 to 16% of ration DM. Recently, Windle and Kung (2013) reported that heifers fed a
spoiling TMR consumed less DM than those fed a fresh total mixed ration (TMR).
Table 1 Some factors that may make silages more prone to aerobic spoilage
Factor
Effects
High sugar content or high
natural population of yeasts
Yeasts use sugars as energy sources
during fermentation
High DM content
a) High DM restricts fermentation
and reduces acid production that
could increase yeast survival during
ensiling
b) High DM crops are more
difficult to pack and allow
infiltration of air into the mass
Poor pack density/porosity
Poor feeding management
Examples
a) sugarcane
a)
Lucerne ensiled > 45 to
50% DM
b)
Maize silage ensiled >
40% DM
Allows penetration of air into the
silage mass
a)
Fill rate too fast
b)
Insufficient compaction
Allows penetration of air into the
silage mass
a)
Slow silage removal
b)
Loose silage
c)
Uneven silage face
d)
Intermediate feeding piles
e)
Silage moved from one
structure to another
Poor management of plastic
and weights
Allows penetration of air into the
silage mass
a)
Torn silo covers
b)
Insufficient weight on plastic
c)
Plastic pulled back too far in
advance
High ambient temperatures
Spoilage organisms grow faster in
warmer weather
More spoilage in the summer than
winter months
Addition of spoiled feeds to
a TMR
Spoiled feeds bring spoilage
organisms to the TMR
Spoiled wet distillers grains
Overly dominant homolactic
acid fermentation
Limited production of organic acids
that have antifungal properties
An extremely dominant homolactic
acid fermentation caused by
microbial inoculation
In contrast, the intake of cows that were fed a TMR containing aerobically spoiling high
moisture corn was unaffected when compared to cows fed fresh corn but the former produced
3.2 kg less milk per cow (Hoffman and Ocker, 1998). When animals consume spoiled silages,
the exact causes of reduced intake and/or performance are not fully understood. In the study
of Whitlock et al. (2000), reduced DM intake probably occurred because of a lower nutrient
digestibility of the silage. However, in the study of Windle and Kung (2013), nutrient
composition, and in vitro digestion of NDF and starch of the diets was very similar and could
not obviously explain the differences in observed intake. One major difference between diets
Proceedings of the 5th Nordic Feed Science Conference
11
Forage conservation and feed processing
was that the fresh TMR contained 5.0 log yeasts/g, whereas the spoiled diet contained 7.8 log
cfu of yeasts/g. Santos et al. (2011) added various levels of a pure culture of I. orientalis to
in vitro ruminal fermentations and reported lower NDF digestibility as the amount of yeast in
the culture increased, suggesting that undesirable spoilage yeasts may have direct effects on
ruminal fermentation. Gerlach et al. (2012) reported negative correlations between
concentration of ethyl lactate and ethanol with DM intake in goats but the strongest negative
effect was related to silage temperature (as difference to ambient). Other variables that may
contribute to depressions of intake include growth of molds in spoiled silage that may result
in production of mycotoxins and effects of microbes or compounds on immune functions.
Organoleptic properties (e.g., taste and smell) of spoiled feeds on intake have not been well
studied. In addition to negative effects on animal performance, spoiled silages also
potentially present a contaminant to the environment if the feed is spoiled to the extent that
must be discarded.
Improving the aerobic stability of silages through management
Filling silos quickly with sufficient pack weight to maximize silage density and minimize
porosity can minimize oxygen in a silo. After filling, silage should be covered with plastic as
soon as possible and weighted down with tires (tires should be touching) or gravel bags to
exclude air. The return on investment (labor and plastic) is extremely high for covering bunk
and pile silos. Oxygen barrier plastics with low transmission rates for oxygen appear to be
useful in minimizing the loss of nutrients at the silage/plastic interface (Borreani et al., 2007).
This practice can also reduce the number of yeasts in silages and improve aerobic stability.
Proper management for removal of silage from silos at the feed bunk with the use of
mechanical equipment (e.g., block cutters and silo facers) can help producers to maximize
profits and production. Enough silage should be removed between facings to minimize
aerobic spoilage. Removal of silage should be such to minimize disruption of the silage face
and loose silage on the ground between feedings. Extreme care should be taken to prevent air
from penetrating between the plastic and reaching the silage mass during feed out and storage
and this can be accomplished by stacking tires, or lining gravel bags on the plastic at the
leading edge of the feeding face.
Improving the aerobic stability of silages with additives
Various chemical additives with antifungal properties have been used to enhance aerobic
stability of silages such as organic acids. For example, buffered propionic acid-based
products are commonly used in North America because they are less corrosive and safer to
handle than the straight acid. It is the undissociated (protonated) form of organic acids that is
responsible for their antifungal properties and its prevalence is dependent on pH. This fact
unfortunately means that more acid is needed to be effective in crops that are naturally
limiting in acids from silage fermentation (e.g. crops with more than 40% DM). At the pH of
a standing crop of lucerne (about 6) only about 1% of propionic acid is in the undissociated
form whereas, at a pH of 4.8, about 50% of the acid is undissociated. The undissociated acid
functions by penetrating into microbial cells and disrupting cytosolic functions after
dissociation. Undissociated acids also remain active on the surface of microorganisms and
compete with amino acids for space on active sites of enzymes. Some weak acids may also
alter permeability of microbial cells (Stratford and Anslow, 1998). Application of buffered
propionic acid-based products in North America ranges from about 0.5 to 2 kg/t of wet
forage, depending on the specific situation. In previous studies, we have found that, as
12
Proceedings of the 5th Nordic Feed Science Conference
Forage conservation and feed processing
expected, the effectiveness of propionic acid based additives increases with higher
application rates (Kung et al., 1998; Kung et al., 2000). Potassium sorbate and sodium
benzoate have also been used to improve aerobic stability of maize silages. For example,
treatment with 0.1% potassium sorbate also improved DM recovery and aerobic stability and
lowered the final concentration of ethanol in maize silage (Teller et al., 2012). Knicky and
Spörndly (2011) reported that an additive containing sodium benzoate, potassium sorbate,
and sodium nitrate improved the aerobic stability of a variety of crops with DM > 35%.
Bacterial inoculants, based on homofermentative lactic acid bacteria are commonly added to
silages to improve fermentation and increase DM and energy recovery. However, most of
these inoculants are not very effective in inhibiting the growth of yeasts because they tend to
maximize the production of lactic acid (poor antifungal activity) and decrease the
accumulation of other organic acids that have good antifungal activity. Muck and Kung
(1997) summarized the literature and found that treatment with classical homolactic acidbased inoculants improved aerobic stability about one third of the time, had no effect about
one third of the time but made aerobic stability worse about one third of the time.
Lactobacillus buchneri, an obligate heterolactic acid bacterium, has been used as a silage
inoculant to specifically enhance aerobic stability by converting moderate amounts of lactic
acid to acetic acid in a variety of silages from maize, sorghum, barley, lucerne, ryegrass,
orchard grass, etc. (Dreihuis et al., 1999a, Kung and Ranjit, 2001). Concerns relative to the
potential of large DM losses from silages treated with L. buchneri, because of its heterolactic
nature, have not been substantiated (Kleinschmit and Kung, 2006). Although, some have
suggested that high levels of acetic acid in silages may depress intake, studies have shown
that ruminants fed silages treated with L. buchneri consume similar amounts of DM as those
fed untreated silages (Dreihuis et al., 1999b, Kung et al., 2003).
Conclusions
Aerobically spoiled silage is undesirable because of losses in nutrients and potential negative
effects on animal performance and health. Good management and the use of various additives
can help to minimize the incidence of aerobically spoiled silage.
References
Borreani, G., Tabacco, E. & Cavallarin, L., 2007. A new oxygen barrier film reduces aerobic
deterioration in farm-scale corn silage. J. Dairy Sci., 90, 4701-4706.
Driehuis, F., Oude Elferink, S.J.W.H. & Spolestra, S.F., 1999a. Anaerobic lactic acid
degradation during ensilage of whole crop maize inoculated with Lactobacillus buchneri
inhibits yeast growth and improves aerobic stability. J. Appl. Microbiol., 87, 583-594.
Driehuis, F., Oude Elferink, S.J.W.H. & Van Wikselaar, P.G., 1999b. Lactobacillus buchneri
improves the aerobic stability of laboratory and farm scale whole crop maize silage but does
not affect feed intake and milk production of dairy cows. Pages 264-265 in Proc. XII Intl.
Silage Conf. Uppsala, Sweden.
Gerlach, K., Weiß, K., Roß, F., Büscher, W. & Südekum, K.-H., 2013. Changes in maize
silage fermentation products during aerobic deterioration and its impact on feed intake by
goats. Agri. Fd. Sci., 22, 168–181.
Hoffman, P. C. & Ocker, S. M., 1997. Quantification of milk yield losses associated with
feeding aerobically unstable high moisture corn. J. Dairy Sci. 80 (Suppl. 1), 234 (Abstr.).
Proceedings of the 5th Nordic Feed Science Conference
13
Forage conservation and feed processing
Inglis, G. D., Yanke, L.J., Kawchuk, L.M. & McAllister, T.A., 1999. The influence of
bacterial inoculants on the microbial ecology of aerobic spoilage of barley silage. Can. J.
Microbiol., 45, 77-87.
Kleinschmit, D. H. & Kung, L., Jr., 2006. A meta-analysis of the effects of Lactobacillus
buchneri on the fermentation and aerobic stability of corn, grass and small grain silages. J.
Dairy Sci., 89, 4005-4013.
Knicky, M. & Spörndly, R., 2011. The ensiling capability of a mixture of sodium benzoate,
potassium sorbate, and sodium nitrite. J. Dairy Sci., 94, 824–831.
Kung, L., Jr., & Ranjit, N.K., 2001. The effect of Lactobacillus buchneri and other additives
on the fermentation and aerobic stability of barley silage. J. Dairy Sci., 84, 1149-1155.
Kung, L., Jr., Taylor, C. C., Lynch, M.P. & Neylon, J.M., 2003. The effect of treating
alfalfa with Lactobacillus buchneri 40788 on silage fermentation, aerobic stability, and
nutritive value for lactating dairy cows. J. Dairy Sci., 86, 336-343.
McDonald, P., Henderson, N. & S. Heron, J. E. (Eds.), 1991. The Biochemistry of Silage.
Second Edition. Chalcombe Publications, Bucks, England.
Muck, R. E. & Kung, L., Jr., 1997. Effect of silage additives on ensiling. Proceedings of the
conference on Silage: Field to feedbunk. North American Conference Hershey, PA.
NRAES-99.
Santos, M. C., Lock, A.L., Mechor, G.D. & Kung, L., Jr., 2011. Spoilage yeasts in silage
have the potential to directly impact rumen fermentation. J. Dairy Sci. 94 (E-Suppl. 1), 207
(Abstr.).
Spoelstra, S. F., Courtin, M. G. & Van Beers, J.A.C., 1988. Acetic acid bacteria can initiate
aerobic deterioration of whole crop maize silage. J. Agric. Sci., Cambridge., 111, 127-132.
Stratford, M. & Anslow, P.A., 1998. Evidence that sorbic acid does not inhibit yeast as a
classic ‘weak acid preservative’. Lett. Appl. Microb., 27, 203–206.
Teller, R. S., Schmidt, R.J., Whitlow, L.W. & Kung, L., Jr., 2012. Effect of physical damage
to ears of corn before harvest and treatment with various additives on the concentration of
mycotoxins, silage fermentation, and aerobic stability of corn silage. J. Dairy Sci., 95, 1428436.
Whitlock, L. A., Wistuba, T.J., Seifers, M.K., Pope, R.V. & Bolsen, K.K., 2000. Effect of
level of surface-spoiled silage on the nutritive value of corn silage diets. J. Dairy Sci. 83
(Suppl. 1), 110 (Abstr.).
Windle, M. C. & Kung, L., Jr., 2013. The effect of a feed additive on the feeding value of a
silage-based TMR exposed to air. J. Dairy Sci. 91(E-Suppl. 1), 16 (Abstr.).
Woolford, M. K., 1975. Microbiological screening of the straight chain fatty acids (C1-C12) as
potential silage additives. J. Sci. Food Agric., 26, 219-228.
14
Proceedings of the 5th Nordic Feed Science Conference
Forage conservation and feed processing
Aerobic exposure of silages – effects on chemical composition and preference by goats
K. Gerlach & K.-H. Südekum
Institute of Animal Science, University of Bonn, Endenicher Allee 15, 53115 Bonn, Germany
Correspondence: [email protected]
Introduction
Whole-crop maize, grass and legume silages play a significant role in the nutrition of
ruminants world-wide. After silo opening, most silages are exposed to air for several days
(silo face, fodder mix waggon, feeding trough) until being consumed by the animal. However, aerobic stability is often shorter than this interval and as a consequence, aerobic
deterioration processes might have taken place before consumption. There is a lack of data
concerning the impact of aerobic spoilage related changes on dry matter intake (DMI) and
preference by ruminants.
Only few experiments focused on the impact of aerobic exposure of silages on DMI. A
possible reason might be the difficulty of keeping silages at specific stages of deterioration
for repeated use during feeding trials. As frost before harvest can have a major impact on
silage digestibility (St-Pierre et al., 1983), freeze-thaw processing of silages for experimental
purposes may also lead to significant changes in silage structure, digestibility and rate of
volatilisation of fermentation products, especially in low-DM substrate. Therefore, freezing is
not the best solution for preserving a given stage of aerobic spoilage of silages for feeding
experiments when sensory characteristics and their influence on DMI are being evaluated.
Sensory characteristics, especially odour and taste play a significant role in the feeding
behaviour of ruminants (Favreau-Peigné et al., 2013). The perception of odour is a very
complex process with information being processed within a few moments. Most naturally
occurring odours may comprise several hundreds of chemical components. The consequence
is that the odour of a source specifies it but is not chemically identical with it (Persaud et al.,
1996). Efforts to reproduce or mimic the complex process have been made, for example by
creating chemosensor/electronic olfactory systems (“electronic nose”, sensor array systems).
A broad range of chemical constituents can be detected in silages, often only by advanced
analytical methods. Increasing numbers of volatile organic compounds that are of special
relevance in ensiled feedstuffs, can be detected and quantified but they may also be closely
interrelated (e.g., ethanol and ethyl esters), and/or complex interactions exist which hamper
clear assignments to production variables. Silage fermentation characteristics have impact on
DMI (e.g., Huhtanen et al., 2002; Eisner et al., 2006 ), but it is very unlikely that the presence
of a single product determines preference for or avoidance against a specific forage in
ruminants. Also microbiological analyses are important and help explaining differences in
spoilage processes between different substrates. Nevertheless, chemical and microbiological
analyses alone might fail in finding points during the course of oxygen ingress that determine
the initiation of spoilage with a concurrent reduction in DMI.
Therefore, preference trials with aerobically exposed silages were carried out to get extended
information and understanding of the animals´ reaction to changing forage qualities.
Materials and Methods
Between 2010 and 2013, a total of 18 preference trials were conducted to determine the
impact of aerobic exposure of different silages on short-time DMI and preference behaviour
by goats. Eight maize silages, eight grass silages (all differing in DM concentration, chop
Proceedings of the 5th Nordic Feed Science Conference
15
Forage conservation and feed processing
length and density) and two lucerne silages (preserved with and without a chemical additive)
were stored aerobically for eight days (d), respectively. Measuring of silage temperature and
sampling for chemical (proximate constituents, fibre fractions and fermentation products), in
vitro (gas production using the Hohenheim gas test) and microbiological analyses (yeasts,
moulds, aerobic mesophilic bacteria and lactic acid bacteria) was conducted in a two-d
interval scheme (d 0, 2, 4, 6 and 8 after silo opening). A detailed description of methods can
be found in Gerlach et al. (2013; 2014). Additionally, in grass and maize silages a
chemosensor system was used for measuring volatile compounds as described by Roß et al.
(2012).
Additionally, samples of the homogenized silage heap from each day of aerobic exposure
(d0, d2, d4, d6 and d8) were stored anaerobically in polyethylene bags (170 µm, 400 mm ×
600 mm; Innovapac, Durach, Germany), which were evacuated and sealed with a singlechamber vacuum packing machine (MAX-F 46; Helmut Boss Verpackungsmaschinen, Bad
Homburg, Germany). A single bag was used for each meal for each goat which was filled
with 1.5 - 1.7 kg silage (requirements + buffer = 60 bags/d for each treatment). Bags were
stored in a dark, dry and cool room (15°C) until used in the preference trial. Storage time of
the silages in the vacuum bags ranged from 1 to 26 d depending on the day when fed.
In 18 trials with Saanen-type goats (n = 6 for maize silages, n = 5 for grass and lucerne
silages, mean body weight 91 ± 12.4 kg), the impact of aerobic exposure on DMI and
preference was studied by presenting each possible pairwise combination of aerobically
stored forages (d0 - d8), each one lasting 21 d. During the experimental period, each forage
combination was presented for 3 h, always allowing free choice between the treatments. The
experimental design was conducted according to Buntinx et al. (1997). The statistical analysis
was done using the SAS procedure Multidimensional Scaling and by analysis of variance.
Results and Discussion
One of the challenges of the experiment was to keep silage samples at specific stages of
deterioration for their later use in preference trials several days to weeks after the period of
aerobic exposure. For this purpose, silages were stored completely airtight in vacuum-sealed
polyethylene-bags, as previously proposed by Pippard et al. (1996) as a method for obtaining
and preserving uniform silages for feeding experiments. When storing different grass silages
in evacuated bags for 18 days, no aerobic spoilage-related changes were observed and
therefore, the methodology was judged as being suitable for preserving silages for use in
trials of the type reported in this publication.
Aerobic stability and degree of changes in fermentation products differed remarkably
between forages. Goats strongly differentiated between aerobically exposed silages, but with
different magnitudes between silage types. In maize silages, avoidance of silages began at the
d4 of exposure, in lucerne silages even earlier. In grass silages, the impact of oxygen was
much less pronounced. When comparing d0- and d8-silages, the mean decline of DMI was
53, 35 and 61% for maize, grass and lucerne silages, respectively.
Although silages were analysed for a wide range of chemical components including more
than 20 fermentation products, it was difficult to find substances being responsible for the
decrease in preference. No fermentation product exhibited a close relation to DMI, i.e. a
correlation coefficient > 0.6. Consequently, it appears from the experiments that assigning a
decline in feed intake to single constituents is difficult if not futile; often the impact of
16
Proceedings of the 5th Nordic Feed Science Conference
Forage conservation and feed processing
Table 1 Dry matter intake (g/3 h) of selected silages with different lengths of aerobic exposure (0-8 days)
shown by goats, n=25/30
Length of aerobic exposure (days)
Substrate
0
2
4
6
8
Maize silage
715 a
657 a
467 b
444 b
256 c
Grass silage
566 a
561 a
630 a
509 a
272 b
Lucerne silage (untreated)
672 a
569 a
427 b
392 b
223 c
Means within rows bearing different superscripts (a, b, c and d) differ (P < 0.05).
individual compounds or their mixtures on preference is still unclear. Huhtanen et al. (2002)
proposed that eventually a complex of fermentation products or still unidentified compounds
might be responsible for effects on intake. The difficulty of identifying single constituents or
understanding chemical reactions being responsible for preference or avoidance of a feed is
emphasized by the fact that silage temperature expressed as difference to ambient
temperature (ΔT) was by far the best single predictor in maize silages. The strong negative
correlation between ΔT and DMI underlines its utiliability for practical silage management
and the suitability of recommendations given by Spiekers et al. (2009).
One of the olfactory systems trying to reproduce the process of odour perception by animals
was also evaluated in our grass and maize silage studies (Roß et al., 2012). Results
demonstrated the possibility of objectively measuring forage quality with the help of sensor
array systems under laboratory conditions. Nevertheless, it was difficult to guarantee constant
measurement conditions and to clearly differentiate between silages. Furthermore, sensor
array systems are known to respond very sensitively to temperature and humidity changes
and signal drift over time and have only moderate reproducibility and high costs (Sankaran et
al., 2012), which might impede the use of sensor arrays directly at the silo face. The fact that
temperature showed a much closer correlation with DMI underlines the need for further
modifications concerning the suitability for daily use (Gerlach, 2013). The taste, which also
strongly influences feeding behavior (Favreau-Peigné et al., 2013), is at least partly neglected
when using olfactory systems.
With the presented methodology, the impact of aerobic exposure of different silages on
chemical characteristics and changes in DMI and preference behaviour can be characterised.
The design of the preference trials lead to a sufficient number of observations in a
manageable interval of time, therefore giving the possibility of using different forages and
conducting several trials successively. The pairwise presentation of each possible
combination of feeds allows the identification of even small differences in forage
characteristics influencing DMI and preference.
The number of fermentation products and other silage characteristics that can be analysed and
detected will likely continue to increase in the future; on the other hand the animal which
perceives a multiple of these compounds by sensory impression and its feedback (shown by
increased or decreased preference and intake) should not be neglected. Sensory
characteristics of feedstuffs have a more important role than generally considered in the
feeding behaviour of domestic ruminants (Favreau-Peigné et al., 2013). It is worth putting the
animal itself and its preference behaviour in focus in further investigations, as it is the only
possibility of covering all silage characteristics (sensory and post-ingestive feedback) the
animals perceive.
Proceedings of the 5th Nordic Feed Science Conference
17
Forage conservation and feed processing
References
Buntinx, S.E., Pond, K.R., Fisher, D.S. & Burns, J.C., 1997. The utilization of
multidimensional scaling to identify forage characteristics associated with preference in
sheep. J. Anim. Sci. 75, 1641-1650.
Eisner, I., Südekum, K.-H. & Kirchhof, S., 2006. Beziehungen zwischen
Fermentationscharakteristika von Silagen und der Futteraufnahme von Milchkühen
[Relationships between silage fermentation characteristics and feed intake by dairy cows].
Übers. Tierernährg. 34, 197-221 (in German with English summary).
Favreau-Peigné, A., Baumont, R. & Ginane, C., 2013. Food sensory characteristics: their
unconsidered roles in the feeding behaviour of domestic ruminants. Animal 7, 806-813.
Gerlach, K., 2013. The aerobic deterioration of silages as estimated from chemical
composition and dietary choice by goats. Dissertation, University of Bonn, 97 pp.
Gerlach, K., Roß, F., Weiß, K., Büscher, W. & Südekum, K.-H., 2013. Changes in maize
silage fermentation products during aerobic deterioration and effects on dry matter intake by
goats. Agric. Food Sci. 22, 168-181.
Gerlach, K., Roß, F., Weiß, K., Büscher, W., Südekum, K.H. & 2014. Aerobic exposure of
grass silages and its impact on dry matter intake and preference by goats. Small Rumin. Res.
117, 131-141.
Huhtanen, P., Khalili, H., Nousiainen, J.I., Rinne, M., Jaakkola, S., Heikkila, T. &
Nousiainen, J., 2002. Prediction of the relative intake potential of grass silage by dairy cows.
Livest. Prod. Sci. 73, 111-130.
Persaud, K.C., Khaffaf, S.M., Payne, J.S., Pisanelli, A.M., Lee, D.-H. & Byun, H.-G., 1996.
Sensor array techniques for mimicking the mammalian olfactory system. Sens. Actuators B
Chem. 36, 267-273.
Pippard, C.J., Porter, M.G., Steen, R.W.J., Gordon, F.J., Mayne, C.S., Poots, R.E., Unsworth,
E.F. & Kilpatrick, D.J., 1996. A method for obtaining and storing uniform silage for feeding
experiments. Anim. Feed Sci. Technol. 57, 87-95.
Roß, F., Boeker, P., Büscher, W., Gerlach, K., Haas, T., Maack, C. & Südekum, K.-H., 2012.
A chemosensor system for assessment of silage quality. Proc. XVI Int. Silage Conf.,
Hämeenlinna, Finland, pp. 111-112.
Sankaran, S., Khot, L.R., Panigrahi, S. & 2012. Biology and applications of olfactory sensing
system: A review. Sens. Actuators B Chem. 171-172, 1-17.
Spiekers, H., Potthast, V. & Nussbaum, H., 2009. Erfolgreiche Milchviehfütterung, DLGVerlags-GmbH, Frankfurt am Main, Germany.
St-Pierre, N.R., Bouchard, R., St-Laurent, G.J., Vinet, C. & Roy, G.L., 1983. Effects of stage
of maturity and frost on nutritive value of corn silage for lactating dairy cows. J. Dairy Sci.
66, 1466-1473.
18
Proceedings of the 5th Nordic Feed Science Conference
Forage conservation and feed processing
Effects of particle size and chemical additives on fermentation and aerobic stability of
grass-clover silage
E. Nadeau1 & H. Auerbach2
1
Swedish University of Agricultural Sciences (SLU), Department of Animal Environment and
Health, 532 23 Skara, Sweden. 2 ADDCON EUROPE GmbH, Parsevalstrasse 6a, 06749
Bitterfeld-Wolfen, Germany
Correspondence: [email protected]
Introduction
Particle size of ensiled forages affects fermentation profile (McEniry et al., 2008) and forage
particle size per se affects intake and performance of cattle and sheep (Rustas and Nadeau,
2011; Helander et al., 2014). Nadeau et al. (2012) observed positive effects on fermentation
in long and chopped grass silage by chemical silage additives. However, as there is still
lacking comprehensive knowledge of the influence of additives on the fermentation of
Clostridium-infected forages differing in particle size, this study aimed at investigating the
effects of chopping and additive use on the fermentation characteristics of grass-clover forage
containing Clostridium spores at ensiling.
Materials and Methods
A second cut ley of grass -clover (50%/50% of dry matter (DM)) ley was mowed on July 17,
2012, wilted to 30% DM and subsequently put through a commercial baler (Deutz-Fahr 160).
At harvest, the forage contained 361 g neutral detergent fibre, 130 g water-soluble
carbohydrates (WSC) and 130 g crude protein per kg DM. The in vitro organic matter
digestibility was 857 g/kg of organic matter. The epiphytic lactic acid bacteria (LAB) count
was 300000 cfu/g fresh herbage, and clostridia were below the limit of detection of 10
MPN/g. Half of the material was left unchopped (long, 180 mm particle size) whereas the
remaining portion was chopped to 17 mm particle size in a laboratory chopper before being
ensiled. Soil containing Clostridium spores was added at 50 g per kg forage to give a
theoretical inoculation rate of 90 MPN/g fresh matter to all the treatments, except for one
treatment that served as the uninoculated control (UIC). Prior to ensiling of the forage
contaminated with soil, the following salt-based chemical additives were applied: KOFASIL
LIQUID (24.5% sodium nitrite; 16.4% hexamine, 2.5 mL/kg; KOFASIL LIQUID Plus
(24.5% sodium nitrite, 5.0% sodium benzoate, 5.0% potassium sorbate, 2.5 mL/kg);
KOFASIL LP (20.2% sodium nitrite, 13.5% hexamine, 5.0% sodium benzoate, 2.5 mL/kg),
KOFASIL ULTRA K (16.5% sodium nitrite, 11.0% hexamine, 8.1% potassium sorbate, 2.2%
sodium benzoate, 0.8% sodium propionate, 2.5 mL/kg); SAFESIL (18.0% sodium benzoate,
7.4% potassium sorbate, 5.0% sodium nitrite, 4.0 mL/kg). Additional treatments were tested
by using the following acid-based additives: GrasAAT SX (40.0% formic acid, 20.0%
sodium formate, 20.0% propionic acid, 1.0% benzoic acid, 1.0% sorbic acid, 4.0 mL/kg),
PROMYR XR680 (48.8% formic acid, 18.4% propionic acid, 6.1% sodium, 4.0 mL/kg). The
additives were diluted with water and applied manually by spraying at 10 mL/kg forage. A
control infected with Clostridium spores (C) and the UIC received 10 mL of water per kg
forage. Silages of the nine treatments for each particle size were prepared in triplicates and
stored in 1.5-L glass jars at 25 °C for 173 days.
The loss of DM during fermentation was calculated according to Weissbach (2005).
Fermentation products were analyzed by HPLC and GC and their concentrations were based
on DM, which was corrected for the loss of volatiles during drying (Weiss and Kaiser, 1995;
Proceedings of the 5th Nordic Feed Science Conference
19
Forage conservation and feed processing
Weiss, 2001; Weissbach and Strubelt, 2008). Ammonia-N concentration was determined
colorimetrically by Scalar (CFSA) based on the Berthelot reaction and the content of
ammonia-N in silages treated with KOFASIL LIQUID, KOFASIL LIQUID Plus, KOFASIL
LP, KOFASIL ULTRA K and SAFESIL were corrected for the amount of N derived from
hexamine (100%) and NO2 (90%) based on previous studies (Zwierz and Weissbach, 1989;
Knicky and Lingvall, 2004). Water soluble carbohydrates (WSC) were determined by the
anthrone method according to Lengerken and Zimmermann (1991). Aerobic stability of the
silages was measured for 20.4 days by wireless temperature loggers (Tinytag Talk 4014,
Gemini, Chichester, UK) and expressed as the number of days before reaching a temperature
of 2°C above ambient (Honig, 1990). Data were analysed as a completely randomized design
in PROC GLM of SAS 9.3, with additive treatment and particle size as fixed factors, using
three replicates per treatment. As some additives had similar effects on studied variables,
contrasts were made from the nine treatments within each particle size. The contrasts used
were UIC vs. C, the mean of the salts vs. C, the mean of the acids vs. C and the mean of the
salts vs. the mean of the acids. To avoid the risk for mass significance, the non-significance
level was lowered to P > 0.01 in the comparison of the means in the contrasts.
Results and Discussion
Table 1 Effects of particle size on fermentation and aerobic stability of grass-clover silages (n = 27).
Parameter
Particle size
SEM
P - value
Chopped
Long
30.8
30.9
0.13
ns
DM losses (%)
6.8
6.9
0.03
<0.01
WSC (% of DM)
0.7
0.8
0.019
<0.01
4.16
4.13
0.005
<0.001
7.5
6.9
0.06
<0.001
Lactic acid (% of DM)
9.79
9.49
0.114
ns
Acetic acid (% of DM)
3.69
3.52
0.059
<0.05
Butyric acid (% of DM)
0.05
0.05
0.01
ns
Ethanol (% of DM)
0.38
0.34
0.007
<0.001
Propanol (% of DM)
0.63
0.56
0.013
<0.001
Aerobic stability (days)
18.4
19.3
0.12
<0.001
DM (%)
pH
NH3-N (% total N)
Even though there were significant interactions between additive treatment and particle size
for some of the variables studied, the differences were not significant between particle sizes
within additive treatment. Thus, the main effect of particle size across additive treatments is
presented. Particle size had no effect on lactic acid and butyric acid concentrations (Table 1).
All other tested parameters showed response to mechanical treatment but the magnitude was
of minor relevance in terms of silage quality. When averaged over treatments, the silages
showed good fermentation characteristics and very high stability upon exposure to air
substantiating findings for grass silages by Nadeau et al. (2012).
The results presented in tables 2 and 3 demonstrate that the forage which was not inoculated
with clostridia fermented well and had low DM losses regardless of particle size. This
supports findings by Weissbach and Honig (1996) who set the critical value for good
fermentation quality of silages from forages to a minimum of 100,000 epiphytic LAB/g,
20
Proceedings of the 5th Nordic Feed Science Conference
Forage conservation and feed processing
which is lower than 300,000 epiphytic LAB found per g forage in this experiment. However,
inoculation of the herbage with Clostridium spores resulted in butyric acid formation as
observed by Kaiser et al. (2002). These authors demonstrated that not only the typically used
parameters characterizing the fermentability of the crop – DM, ratio of water-soluble
carbohydrates and buffering capacity– affect the quality of the fermentation process, but also
the clostridia level, as well as the nitrate content.
Table 2 Effects of Clostridium inoculation and of chemical silage additives on fermentation, dry matter (DM)
losses and aerobic stability of chopped grass-clover silage.
Parameter
P contrasts
Treatments
UIC1
C2
Salts
Acids
UIC vs
Salts vs
Acids vs
Salts vs
C
C
C
Acids
ns
ns
ns
ns
DM (%)
29.0
30.5
30.9
31.6
DM losses (%)
6.1
8.1
7.0
6.2
<0.001
<0.001
<0.001
<0.001
pH
4.08
4.39
4.18
4.02
<0.001
<0.001
<0.001
<0.001
WSC (% of DM)
2.46
0.32
0.43
0.70
<0.001
ns
<0.001
<0.001
NH3-N (% of total-N)
8.1
6.8
7.5
<0.001
<0.001
<0.001
<0.001
Lactic acid (% of DM)
8.66
8.03
10.01
10.66
ns
<0.001
<0.001
ns
Acetic acid (% of DM)
2.75
4.14
4.08
2.97
<0.001
ns
<0.001
<0.001
Propionic acid (% of DM)
0.02
0.25
0.12
0.28
<0.001
<0.001
ns
<0.001
Butyric acid (% of DM)3
0.00
0.43
0.00
0.00
<0.001
<0.001
<0.001
ns
Ethanol (% of DM)
0.59
0.85
0.32
0.21
<0.001
<0.001
<0.001
<0.001
Propanol (% of DM)
0.01
0.86
0.77
0.46
<0.001
ns
<0.001
<0.001
<0.001
ns
<0.001
Aerobic stability (days)
1
10.6
10.5
20.4
18.9
20.0
2
<0.001
3
UIC = uninoculated control, C = control inoculated with Clostridium spores, below detection limit of 0.01%
of fresh matter.
Higher clostridia contamination increases the risk of undesired clostridia fermentation.
Despite butyric acid formation, ammonia levels were still acceptable and the pH was low.
This can be explained by the activity of saccharolytic clostridia in the early stages of
fermentation, which is associated with only limited proteolysis (Kaiser et al., 1997)
compared with the activity of proteolytic butyric acid bacteria at later stages of fermentation
(Weiss and Kaiser, 2002). In addition to the presence of butyric acid, inoculated silages also
contained higher acetic acid concentrations, thereby rendering them more stable upon
exposure to air (Tables 2 and 3; Auerbach et al., 2013; Nadeau and Auerbach, 2013).
Obviously, inoculation with clostridia also had an impact on other metabolic pathways during
the fermentation process as indicated by higher ethanol, propanol and propionic acid contents
as compared to the UIC. Although 1,2-propanediol was not found in the silage, it most likely
was produced by epiphytic or soil-originating Lactobacillus buchneri bacteria in
combination with acetic acid from anaerobically degraded lactic acid. Further on, 1,2propanediol was utilized by Lactobacillus diolivorans, to form 1-propanol and propionic acid
(Oude-Elferink et al., 2001; Krooneman et al., 2002). These changes in metabolic pathways
led to greater losses of WSC and DM in the C than in the UIC treatment.
Proceedings of the 5th Nordic Feed Science Conference
21
Forage conservation and feed processing
Table 3 Effects of Clostridium inoculation and of chemical silage additives on fermentation, dry-matter (DM)
losses and aerobic stability of long grass-clover silage.
Parameter
P contrasts
Treatments
1
UIC
DM (%)
2
C
Salts
Acids
UIC vs
Salts vs
Acids vs
Salts vs
C
C
C
Acids
28.3
31.1
31.1
31.5
<0.001
ns
ns
ns
DM losses (%)
6.2
8.4
7.0
6.4
<0.001
<0.001
<0.001
<0.001
WSC (% of DM)
2.02
0.57
0.67
0.62 <0.001
ns
ns
ns
pH
4.13
4.27
4.13
4.04 <0.001
<0.001
<0.001
<0.001
NH3-N (% of total-N)
8.2
9.8
6.0
7.1
<0.001
<0.001
<0.001
Lactic acid (% of DM)
8.18
7.81
10.26
9.05 ns
<0.001
<0.01
<0.001
Acetic acid (% of DM)
2.60
4.49
3.76
2.90 <0.001
<0.001
<0.001
<0.001
0.01
0.17
0.10
0.22 <0.001
<0.001
<0.01
<0.001
0.00
0.46
0.00
0.00 <0.001
<0.001
<0.001
ns
Ethanol (% of DM)
0.48
0.71
0.29
0.19 <0.001
<0.001
<0.001
<0.001
Propanol (% of DM)
0.03
0.88
0.66
0.40 <0.001
<0.001
<0.001
<0.001
ns
ns
ns
Propionic acid (% of DM)
Butyric acid (% of DM)
3
Aerobic stability (days)
1
10.5
2
20.4
20.4
20.4
<0.001
<0.001
3
UIC = uninoculated control, C = control inoculated with Clostridium spores, below detection limit of 0.01%
of fresh matter.
Both the salts and the acids inhibited the growth of clostridia as indicated by no butyric acid
formation and an increased lactic acid production while the acetic acid and ethanol
concentrations were decreased, indicating a more homolactic fermentation. This resulted in
lower pH in the silages treated with the chemical additives compared with the C treatment
(Tables 2 and 3). The more extensive acidification in the treated silages resulted in less
proteolytic activity with a lower ammonia-N concentration in the additive-treated silages
compared to the C treatment. The improved fermentation resulted in less losses of DM in the
silages treated with the chemical additives compared to the C treatment (Tables 2 and 3).
Similar positive effects on silage fermentation and on the extent of proteolysis by chemical
additives have previously been reported (Auerbach et al., 2012). The aerobic stability was
high in all silages.
The acid-treated silages had lower concentrations of acetic acid and ethanol and lower pH
and DM losses than the salts in both chopped and long silages (Tables 2 and 3). However, the
lactic acid concentration was higher in the silage treated with the salts compared with the acid
treatment when ensiled as long forage. Also, the salts decreased the proteolysis of protein as
indicated with a lower ammonia-N concentration in both chopped and long silages treated
with the salts compared to the acids which is in agreement with results by Nadeau et al.
(2012).
Conclusions
The influence of particle size on fermentation and aerobic stability was less pronounced than
that of chemical additives. As contamination of forage by Clostridium-containing soil cannot
always be avoided under practical farming conditions, strategic use of chemical additives is
22
Proceedings of the 5th Nordic Feed Science Conference
Forage conservation and feed processing
strongly advised in order to alleviate the adverse effects on fermentation characteristics by
Clostridium bacteria.
Acknowledgements
The authors are deeply indebted to Agroväst, Skara, Sweden for funding this experiment and
to Annika Arnesson, SLU, Skara, for technical support during the trial. Dr. Kirsten Weiss,
Humboldt University, Berlin, Germany, is acknowledged for conducting the analyses of
fermentation products in the silages.
References
Auerbach, H., Nadeau, E. & Weiss, K., 2012. Benefits of using silage additives, in:
Auerbach, H., Lückstädt, C., Weissbach, F. (Eds.), The future of silage preservation.
Proceedings of the 1st International Silage Summit, 12 November, Leipzig, Germany, pp. 75144. ISBN: 978-0-9573721-6-0.
Auerbach, H., Weiss, K., Theobald, P. & Nadeau, E., 2013. Effect of inoculant type on dry
matter losses, fermentation pattern, yeast count and aerobic stability of green rye silages, in:
Mair, C., Kraft, M., Wetscherek, W., Schedle, K. (Eds.), 12. BOKU-Symposium
Tiernernährung. Universität für Bodenkultur, 11 April, Wien, Austria. pp. 179-185.
Helander, C., Nørgaard, P., Arnesson, A., Nadeau, E. & 2014. Effects of chopping grass
silage and of mixing silage with concentrate on feed intake and performance in pregnant and
lactating ewes and in growing lambs. Small Rumin. Res.116, 78-87.
Honig, H., 1990. Evaluation of aerobic stability. Grass Forage Reports, Special issue 3, 7682.
Kaiser, E., Weiss, K. & Zimmer, J., 1997. Zum Gärungsverlauf bei der Silierung bei
nitratarmem Grünfutter. 1. Mitteilung: Gärungsverlauf in unbehandeltem Grünfutter (On the
course of fermentation in silages from low-nitrate forages. First part: The course of
fermentation of untreated forages). Arch. Tierern. 50, 87-102.
Kaiser, E., Weiss, K. & Polip, I., 2002. A new concept for the estimation of the ensiling
potential of forages, in: Gechie, L.M., Thomas, C. (Eds.), Proc. 13th Int. Silage Conf., 11-13
Sept., Auchincruive, Scotland, pp. 344-357.
Knický, M. & Lingvall, P., 2004. Ensiling of high wilted grass-clover mixture by use of
different additives to improve quality. Acta Agric. Scand., Sect. A, Animal Sci. 54, 197-205.
Krooneman, J., Faber, F., Alderkamp, A. C., Oude Elferink, S.J. H.W., Driehuis, F.,
Cleenwerck, I., Swings, J., Gottschal, J.C. & Vancanney, M., 2002. Lactobacillus diolivorans
sp. nov., a 1,2-propanediol-degrading bacterium isolated from aerobically stable maize silage.
Intern. J. Syst. Evolution. Microbiol. 52, 639-646.
Lengerken, von, J., Zimmermann, K., 1991. Handbuch Futtermittelprüfung (Handbook feed
evaluation) . Deutscher Landwirtschaftsverlag Berlin, 1. Auflage.
McEniry, J., O´Kiely, P. Clipson, N.J.W., Forristal, P.D. & Doyle, E.M. 2008. The
microbiological and chemical composition of silage over the course of fermentation in round
bales relative to that of silage made from unchopped and precision-chopped herbage in
laboratory silos. Grass Forage Sci. 63, 407-420.
Proceedings of the 5th Nordic Feed Science Conference
23
Forage conservation and feed processing
Nadeau, E., Arnesson, A. & Auerbach, H., 2012. Effects of additive and particle size on
fermentation characteristics and aerobic stability of grass silage, in: Kuoppala, K., Rinne, M.,
Vanhatalo, A. (Eds.), Proc. XVIth Int. Silage Conf., 2-4 July, MTT Agrifood Research &
University of Helsinki, Hämeenlinna, Finland, pp. 380-381.
Nadeau, E. & Auerbach, H., 2013. Additive type affects dry matter losses, fermentation
pattern, aerobic stability and clostridia counts of baled red clover-grass silage. Article no. 35.
III Int. Symp. Forage Quality Conserv. July 22-23, Campinas, SP, Brazil.
Oude Elferink, S.J.W.H., Krooneman, J., Gottschal, J.C., Spoelstra, S.F., Faber, F. &
Driehuis, F., 2001. Anaerobic conversion of lactic acid to acetic acid and1,2-propanediol by
Lactobacillus buchneri. Appl. Environm. Microbiol. 67(1), 125-132.
Rustas, B.O. & Nadeau, E., 2011. Chopping of whole-crop barley silage improves intake and
live weight gain of young dairy steers. Livest. Sci. 141, 80-84.
Weiss, K., 2001. Gärungsverlauf und Gärqualität von Silagen aus nitratarmem Grünfutter
(Course of fermentation and fermentation quality of silages made from low-nitrate forages).
Dissertation. Humboldt-Universität zu Berlin.
Weiss, K. & Kaiser, E., 1995. Milchsäurebestimmung in Silageextrakten mit Hilfe der HPLC
(Determination of lactic acid in silage extracts by HPLC). Das wirtschaftseigene Futter. 41,
69-80.
Weiss, K. & Kaiser, E. 2002. Fermentation patterns in silages from herbage low in nitrate, in:
Gechie, L.M., Thomas, C. (Eds.), Proc. 13th Int. Silage Conf., 11-13 Sept., Auchincruive,
Scotland, pp. 414-415.
Weissbach, F., 2005. A simple method for the correction of fermentation losses measured in
laboratory silos, in: Park, R.S., Stronge, M.D. (Eds.), Silage production and utilization, Proc.
14th Int. Silage Conf., July, Belfast, Northern Ireland, p. 278.
Weissbach, F. & Strubelt, C., 2008. Correcting the dry matter content of grass silages as
substrate for biogas production. Landtechnik 63(4), 210-211.
Weissbach, F. & Honig, H., 1996. Über die Vorhersage und Steuerung des Gärungsverlaufs
bei der Silierung von Grünfutter aus extensivem Anbau (On the prediction and the control of
the course of fermentation in silages from extensively grown forages). Landbauforschung
Völkenrode, FAL, 46, 10-17.
Zwierz, P.M. & Weissbach, F. 1989. Investigations into preservative residues in silage and
milk, Proc. Int. Symp. Production, evaluation and feeding of silage, 12-16 June, Rostock,
Germany, pp. 185-195.
24
Proceedings of the 5th Nordic Feed Science Conference
Forage conservation and feed processing
Effects of different moisture contents and additives on the quality of baled oat silage
G. Q. Zhao*, F.C. Qin, T. Jiao, J.J. Hou & X.D. Song
College of Pratacultural Science, Gansu Agricultural University, Lanzhou 730070, China
Correspondence: [email protected]
Introduction
In recent years, effects of additives on quality of alfalfa, smooth vetch, ryegrass silage have
been studied in China (Ma et al., 2010; Chen et al., 2013; Zhang et al., 2009). Additives such
as lactic acid bacteria (Re, 2009), urea (Zang et al., 2012), corn flour (Guo et al., 2005) and
others have been used in silage making. Besides additives, moisture content of the raw
material is also an important factor affecting silage quality. High water contents could
increase nutrient losses and the risk of badly fermented silage. In contrast, low water contents
might cause substantial nutrient losses due to suboptimal compaction and increased growth
and activity of fungi (Wang et al., 2007).
Oat (Avena sativa L.) is an important feed in high altitude regions of NW China. It is
commonly used to make hay, but frequent rains during autumn make it difficult to get high
quality hay (Liu et al., 2007). Oat in these areas is therefore more suitable to preserve as
silage. Research on either hay or silage making is insufficient (Zhao et al., 2004). There are
no reports on suitable water contents and the use of additives for oat silage. Therefore, the
objectives of this paper was to study the effects of different moisture contents and additives
on the quality of baled oat silage, on the nutritive and fermentation quality and on viable
counts of some microbial groups in oat silage, providing guidance for oat silage making in
this region.
Materials and methods
The experiment was conducted in Xiahe County, which is located in Gannan Tibetan
Autonomous Prefecture (E102°83′, N35°23′, 2517 m altitude).The maximum temperature
was 29.7°C, the minimum -24.1 °C with an annual average temperature of 3.9°C; annual
rainfall was 489 mm. The soil organic matter content was 1.98%, total N 1.18 g/kg, available
nitrogen 60 mg/kg, available phosphorus 57 mg/kg, and soil pH 7.84.
The oat cultivar Longyan No.3 (Avena sativa L.) was provided by the College of Pratacultural
Science, Gansu Agricultural University. The seeding rate was 255 kg/ha. Before planting,
phosphate ammonia was applied at the rate of 225 kg/ha. Corn flour, urea, Synlac Dry
(bacterial inoculant, from Yaxin Biological Bcience and Technology Company, China), and
Sila-Max 200 (bacterial inoculant from Ralco Nutrition Company, USA) were used as
additives.
Oat was harvested at milk stage with a disc mower and was wilted to target moisture contents
of 45-50% (A1) and 65-70% (A2). It was then treated with corn flour (B1), urea (B2), Synlac
Dry (B3), Sila-Max 200 (B4) or baled directly without additives (CK). The amount of
additives were all based on fresh weight of raw material, namely 4%, 0.4%, 0.0002% and
0.00025%, respectively (Table 1). Oat was baled with a round baler and 6 bale replicates per
treatment were produced. Bales weighed approximately 50 kg and were wrapped with 5
layers of stretch film as soon as baled. Bales were sampled 40 d after baling.
Proceedings of the 5th Nordic Feed Science Conference
25
Forage conservation and feed processing
Table 1 Experimental design of the oat silage trial
Moisture content
Additives
Control（CK）
Corn flour（B1）
Urea (B2)
SynlacDry (B3)
Sila-Max200 (B4)
45%-50%（A1）
A1CK
A1B1
A1B2
A1B3
A1B4
65%-70%（A2）
A2CK
A2B1
A2B2
A2B3
A2B4
Before sampling, the stretch film was removed from the bales and 200 g samples were
collected from 3-4 locations of the bale and composited for each bale. Samples were dried at
105°C for 15 min then kept at 65°C until stable weight. After masceration, the samples were
sieved through 40 mesh sieve and sealed in zip-lock bags. Crude protein (CP), neutral
detergent fiber (NDF), acid detergent fiber (ADF), ammonia-N and soluble sugars (WSC)
contents were determined (Yang 1998).
For the detection of organic acids, 20 g silage was mixed with 180 mL deionized water and
kept for 24 h at 4°C, then filtered through 4 layer filter paper, before passing a 0.22 μm
membrane. HPLC (Agilent 1260 and G1321B ultraviolet fluorescence detector) was used to
determine the contents of lactic acid (LA), acetic acid (AA) and acrylic acid (PA) and butyric
acid (BA). Chromatographic conditions: SB-AQ C18 column (4.6 × 250 mm), mobile phase
(methanol): the mobile phase (0.01 mol/L (NH4)2HPO4, pH 2.70) was 3 : 97, velocity was 1
mL/min; detection wave length 210 nm, injection volume 20 μL and temperature was 25°C.
Viable counts were determined by mixing 20 g silage with 180 mL of 0.85% sterilized saline.
A flat gradient dilution method was used to determine the counts of the following microbial
groups: lactic acid bacteria (LAB) were quantitatively determined by using MRS media,
generally aerobic bacteria (Bac) by using defined agar medium and moulds and yeast (M&Y)
by using Tiger red agar medium.
The PC programs SPSS18.0 and SAS JMP10 were used to perform a 2×5 factorial analysis of
variance.
Results and analysis
Moisture contents and additives had both a significant impact on most oat silage parameters
(Table 2). Significant moisture content × additive interactions were observed on all indicators
except Bac. Compared with those parameters of A1 treatment, after adding corn flour (B1),
CP, WSC, LA, M&Y and LAB of A2 treatment did not change much, while NDF, ADF and
NH3-N values were reduced 8.28%, 8.28%, 8.99%, respectively (Table 3, 4 & 5). CP content
increased 15.81%, the LA, AA, PA, Bac and LAB remained stable, but M&Y decreased
13.33%. Synlac Dry had little impact on CP, ADF, AA, PA, M&Y, Bac and LAB, but LA
increased 21.69%; meanwhile, NDF and NH3-N dropped 6.01% and 10.78%, respectively.
After adding Sila-Max 200, CP, LA, AA, PA, M&Y, LAB of A2 treatment maintained stable,
WSC increased while NDF, ADF and M&Y reduced 9.06%%, 18.84% and 14.05%,
respectively, compared with those of A1 treatment . Compared to A1, CK under A2 had a
lower content of NDF and a higher LA (increasing 62.71%). A2 (65%-70%) is more
conducive to oat silage.
Oat silage quality was significantly affected by additives (Tables 3, 4, 5). With the high
moisture level (A2 treatment), CP, NDF and ADF contents of oat silage with corn flour did
not change compared with the control, while WSC increased by 21.22% and LA and AA
26
Proceedings of the 5th Nordic Feed Science Conference
Forage conservation and feed processing
Table 2 F-values of variables including interactions
Moisture content
Additive
Moisture content*Additive
F Value
F Value
F Value
1,4
6.55*
20.06**
3.13*
NDF
1,4
59.35**
3.20*
5.43**
ADF
1,4
30.88**
4.60**
7.72**
WSC
1,4
56.91**
95.46**
167.57**
LA
1,4
2.54
232.36**
29.78**
AA
1,4
41.14**
45.32**
6.83*
NH3-N
1,4
7445.02**
31761.36**
3605.83**
M&Y
1,4
1.05
19.60**
14.15**
Bac
1,4
0.31
18.83**
2.14
LAB
1,4
0.34
18.03**
3.11*
Analyses
df
CP
Note: * and ** indicate significant differences at p<0.05 and p<0.01, respectively
decreased by 53.13% and 30.00%, respectively; microbial groups did not change. Urea
treatment increased CP by 22.78%, ADF increased a little while LA and WSC were reduced
by 32.29% and 10.53%, respectively. Levels of NH3-N and M&Y increased significantly but
no remarkable change was observed on LAB counts. In contrast with the control under A2
treatment, CP, volatile fatty acid, M&Y and Bac did not change after adding Synlac Dry, but
WSC increased by 43.97% and LAB counts increased significantly; ADF and NH3-N were
reduced 3.34% and 61.52%, respectively. Sila-Max 200 increased oat silage CP, WSC, LA
and AA by 11.61%, 19.24%, 40.63% and 43.33%, respectively, compared with the control;
while NDF, ADF and NH3-N decreased 4.77%, 7.59%, and 46.09%; LAB increased
significantly, M&Y and Bac decreased 19.57% and 10.93%, respectively.
Table 3 Nutritional composition of oat silages (mean ± SD)
Item
CP (%DM)
NDF (%DM)
ADF (%DM)
WSC (%DM)
A1CK
9.56±0.01
bc
55.41±0.33
bcd
34.77±1.27
bcd
10.10±0.17
a
A2CK
9.13±0.22
bcd
54.73±0.97
bcde
34.78±0.60
cbd
6.55±0.07
d
A1B1
9.26±0.21
bc
58.71±0.33
a
37.80±0.71
ab
8.31±0.06
c
A2B1
9.04±0.31
bcd
53.85±1.30
cde
33.71±0.91
cd
7.94±0.05
c
A1B2
9.68±0.33
bc
56.45±0.19
bcd
37.75±0.38
ab
7.75±0.20
c
A2B2
11.21±0.58
a
55.28±0.47
bcd
37.28±1.65
abc
5.86±0.05
e
A1B3
7.75±0.09
e
56.06±0.01
abc
35.37±0.97
bcd
7.95±0.8
c
A2B3
8.41±0.40
de
52.69±0.86
de
33.62±0.77
cd
9.43±0.22
b
A1B4
9.84±0.10
abc
57.31±0.24
ab
39.60±0.55
a
6.32±0.07
de
A2B4
10.19±0.00
ab
52.12±0.19
e
32.14±0.63
d
7.81±0.05
c
Note: Different letters in the same column indicate significant difference (P<0.05).
Under A2 moisture treatment, CP content of Sila-Max 200 treatment was similar with that of
urea treatment, 12.72% and 21.17% higher than that of corn flour and Synlac Dry treatment;
Proceedings of the 5th Nordic Feed Science Conference
27
Forage conservation and feed processing
their NDF and ADF contents were the lowest, with the highest LA and AA content;
meanwhile, their M&Y counts were 19.08% and 14.33% higher than the corn flour and the
Synlac Dry treatment. Sila-Max 200 had similar Bac counts with the Synlac Dry treatment,
but 12.59% and 7.86% lower than the corn flour and urea treatment; its LAB counts were
highest, followed by Synlac Dry, both significantly higher than the corn flour treatment
(Table 5).
Table 4 Organic acids and ammonia nitrogen of oat silages (mean ± SD)
Item
LA（%DM）
AA（%DM）
PA（%DM）
NH3-N（%TN）
A1CK
0.59±0.01
de
0.42±0.01
a
0.32±0.01
bc
10.90±0.07
f
A2CK
0.96±0.01
b
0.30±0.02
bc
0.27±0.02
cde
20.87±0.07
c
A1B1
0.52±0.02
ef
0.29±0.01
bcd
0.41±0.02
a
13.12±0.02
d
A2B1
0.45±0.01
f
0.21±0.01
e
0.31±0.02
bcd
11.94±0.26
e
A1B2
0.72±0.02
cd
0.36±0.02
ab
0.34±0.02
abc
38.87±0.11
b
A2B2
0.65±0.01
de
0.29±0.01
bcd
0.37±0.03
ab
68.68±0.14
a
A1B3
0.83±0.02
bc
0.28±0.01
cde
0.23±0.01
de
9.00±0.09
g
A2B3
0.65±0.01
de
0.22±0.01
de
0.23±0.01
de
8.03±0.02
h
A1B4
1.25±0.08
a
0.40±0.02
a
0.23±0.02
de
7.55±0.23
h
A2B4
1.35±0.02
a
0.43±0.01
a
0.18±0.01
e
11.25±0.23
ef
Note: Butyric acid (BA) was not detected in any treatment, therefore not listed.
The moisture content of the fresh forage had a significant impact on silage fermentation. Yan
et al. (1996) found that LAB counts were sensitive to moisture content. Too high or too low
moisture contents are detrimental to silage quality. Generally, suitable moisture content for
oat silage is considered to be around 65%. The results of this study indicate that 65%-70%
moisture content can lead to a better silage quality than 40%-50%. The content of CP and
organic acids of oat silage with higher moisture level (65%-70%) were similar to the one with
the lower level (45%-50%), but NDF and ADF were significantly reduced. On the other
hand, although M&Y under some additive treatments was influenced by moisture content,
LAB counts (approx. 108/g) were always much higher than for M&Y (approx. 103/g). No
significant difference was observed for LAB and Bac counts (Table 5). These results suggest
that the effects of moisture content on oat silage nutrients was not just caused by its effect on
microorganism quantity, it may also have been caused by differences in cell decomposition
rates and dissolved substances, thereby affecting microbial metabolic rate and metabolic
pathways.
The aim of using additives is to create suitable conditions for a favourable fermentation so
that the LAB propagate rapidly, accelerate the fermentation process and thus help to improve
the fermentation quality of silage (Guo et al., 2005). Adding Sila-Max 200 or Synlac Dry
increased LAB counts and improved silage fermentation by reducing pH and improving
fermentation conditions (Driehuis & Vanw, 2000). Meanwhile, large number of LAB
secreted a lot of enzymes to effectively degrade plant tissues which decreased NDF and ADF
levels. The addition of urea and corn flour did not improve this early stage of fermentation,
and did not decrease NDF and ADF values significantly. Our results were not in agreement
with Yang and Zhang’s (2005) results, namely that urea had a dilution effect on degradation
of fresh forage. In contrast, our results showed that adding urea reduced the WSC level and
28
Proceedings of the 5th Nordic Feed Science Conference
Forage conservation and feed processing
Table 5 Microbial counts (log10 cfu/g) of treated oat silages (mean ± SD)
Item
M&Y
Bac
LAB
A1CK
3.19±0.08
ab
4.32±0.09
a
7.70±0.10
e
A2CK
3.27±0.04
a
4.21±0.02
ab
8.05±0.03
cde
A1B1
2.89±0.03
bc
4.19±0.03
a
7.92±0.16
de
A2B1
3.25±0.03
a
4.29±0.07
ab
8.15±0.13
bcde
A1B2
3.00±0.05
ab
4.12±0.01
ab
8.32±0.10
abc
A2B2
2.60±0.15
c
4.07±0.06
ab
8.37±0.15
abcd
A1B3
3.20±0.03
ab
3.77±0.02
cd
8.79±0.06
a
A2B3
3.07±0.04
ab
3.98±0.07
bc
8.61±0.07
ab
A1B4
3.06±0.05
ab
3.56±0.07
d
8.45±0.05
abc
A2B4
2.63±0.04
c
3.75±0.08
cd
8.69±013
a
increased NDF and ADF levels. Urea eventually decomposes to NH3, which dissolves in
water to form ammonium hydroxide and weakens the lignin bonds. But, this reaction may
need a longer period than 40 d to affect silage composition. Meanwhile, the degradation of
WSC leads to an increase in NDF and ADF.
It is generally believed that remaining oxygen at the beginning of the fermentation makes
proteinases retain vitality and decomposes protein into ammonia and amine (Kung et al.,
2003). In addition, under aerobic condition, anaerobic bacteria such as clostridia and
intestinal bacteria decompose amino acid, resulting in increasing protein loss (Herson et al.,
1986). Sila-Max 200 and Synlac Dry treatments significantly increased LAB counts at the
beginning of the fermentation when oxygen was still present. As LAB multiplied, the growth
of other aerobic microbes was significantly inhibited, pH reduced and the content of NH3–N
decreased significantly.
There are more than 20 varieties of LAB which are active during silage fermentation. They
can be grouped into homolactics and heterolactics. Li et al.(2004) believed homolactic
bacteria to be more effective than heterolactics, while Guo’s (2006) study showed the
opposite result. MacDonald et al. (1991) also pointed out that metabolic processes differ
among LAB. Although the applied LAB quantities of Synlac Dry and Sila-Max 200
treatments were not significantly different in our experiment, the latter produced substantially
higher LA , CP and volatile fatty acids (VFA) levels, and lower NDF and ADF levels than
that of Synlac Dry. This may have been due to the different LAB strains contained in Synlac
Dry and Sila-Max 200. As LAB strains have different metabolic pathways during lactic acid
fermentation, choosing the most efficient LAB strains is an important prerequisite to ensure
good fermentation. In addition, silage inoculants containing only one strain may not able to
meet practical needs as multiple LAB strains might be able to.
To achieve an appropriate level of soluble carbohydrates in forages is a prerequisite for good
silage fermentation (Yu et al., 2009). Due to high soluble sugar content and low buffer
capacity, oats can be used to make high-quality silage without any additive. In this study, BA
was not detected in any treatments including the control, indicating that unwanted butyric
acid fermentation was well controlled and fermentation quality was good. However, by using
appropriate additives better silage quality was achieved with lower pH and NH3-N, and
higher LA levels (Re, 2009).
Proceedings of the 5th Nordic Feed Science Conference
29
Forage conservation and feed processing
Conclusions
Moisture content had a significant effect on oat silage quality. At 65%-70% moisture content,
CP, organic acids, M&Y and LAB of additive-treated oat silages were comparable to those at
45%-50% moisture content; while NDF, ADF, WSC and M&Y levels were reduced and
NH3-N increased, which made the wetter the better silage.
Although oat ensiles easily without additives, they could accelerate the fermentation process
and reduce nutrient losses. Compared with the non-bacterial additives such as corn flour and
urea, LAB-based additives silage gave better silage quality and among the two tested, SilaMax 200 gave better results than Synlac Dry.
References
Chen, P. F., Bai, S. Q., Yang, F. Y. et al., 2013. Effects of additive and moisture content on
fermentation quality of smooth vetch silage. Acta Pratacult. Sinica, 22, 80-86.
Driehuis, F. & Vanw, P. G., 2000. The occurrence and prevention of ethanol fermentation in
high-dry-matter grass silage. J. Sci. Food Agric., 80, 711-718.
Guo, X. S., Zhou, H. & Yu, Z., 2006. Effect of different additives on aerobic stability of
silage. China Dairy Cattle 9, 18-20.
Guo, Y., Yang, Q. J. & Wang, H.B., 2005. Effects of different levels of corn flour the nutrient
composition of alfalfa silage. Feed and Husband., 6, 20-22.
Herson, S.E., Edwards, R.A. & McDonald, P., 1986. Changes in the nitrogen components of
gamma irradiated and inoculated ryegrass. J. Sci. Food Agric., 37, 979-985.
Kung, L.Jr., Taylor, C.C., Lynch, M.P. & Neylon J.M., 2003. The effect of treating alfalfa
with Lactobacillus 40788 on silage fermentation, aerobic stability, and nutritive value for
lactating dairy cows. J. Dairy Sci. 86, 336-343.
Li, J.X., Gao, J., Wang, J.H. et al., 2004. Study of lactic acid bacteria biomass in silage
fermentation. Feed. Ind. Mag., 25, 48-51.
Liu Z., Wu, G., Ren, Q. et al., 2007. Sustainable development of animal husbandry based on
oat in alpine grassland area. Pratacult. Sci., 24, 67-69.
Ma, C.H., Xia, Y.J., Han, J., et al., 2010. Effects of different additives on the quality of
alfalfa silage. Acta Pratacult. Sinica, 19, 128-133.
Re, M.N., 2009. Effects of Lactobacillus additives on silage quality of Qing Hai Tibet Plateau
Oats. J. Anhui Agri. Sci. 37, 146- 149.
Wang, L., Li, X.L. & Zhang, X., 2007. The effect of different water contents and additive
mixtures on alfalfa silage. Acta Pratacult. Sinica16, 40-45.
Yan, T.D., Patterson, C. & Gordon, F.J., 1996. The effects of bacteria inoculation of unwilted
and wilted grass silages on silage fermentation, rumen microbial activity, silage nutrient
degradability, and digestibility. Proc. of the 11th Intern. Silage Conf., IGER, Aberystwth, UK.
p. 98-99.
Yang, F.Y. & Zhang, Y.W., 2005. A review on the urea-treated silage. China Dairy Cattle 5,
23-25.
30
Proceedings of the 5th Nordic Feed Science Conference
Forage conservation and feed processing
The effect of silage additive on fermentation quality, clostridia and aerobic stability of a
white lupin-wheat silage
W. König1, M. Lamminen1, K. Weiss3, T.T. Tuomivirta2, S. Sanz Muñoz 2, H. Fritze2, K.
Elo1, L. Puhakka1, A.Vanhatalo1 & S. Jaakkola1
1
Department of Agricultural Sciences, P.O. Box 28, FI-00014 University of Helsinki,
Finland; 2Finnish Forest Research Institute (Metla), Vantaa Unit, P.O. Box 18, FI-01301
Vantaa, Finland; 3)U Berlin, Invalidenstraße 42, 10115 Berlin
Correspondence: [email protected]
Introduction
Bi-crop cultivation has a positive effect on dairy production, farm economy and environment
(Hauggaard-Nielsen et al. 2008). White lupin is one potential legume bi-crop to be used as
whole crop silage because of its high yield (Fraser et al. 2005, Azo et al 2012). Using grain as
bi-crop for legumes improves ensiling quality of legumes, which are considered as difficult to
ensile due to their low dry matter (DM) content and high buffering capacity (McDonald et al.
1991). During the last 20 years, there has been intensive development of technologies to
identify microbial species and estimate their abundancies using polymerase chain reaction
(PCR) (Takeuchi et al. 1997). Information from DNA sequences of 16S gene and its copy
numbers can be used to estimate microbial diversity and abundance in various types of
samples (Bagge et al. 2009, Weimer and Stevenson 2012). The ensiling experiment was
conducted to study the effects of different types of additives on the quality of silage made on
bi-crop differing in white lupine to wheat ratio and maturity.
Materials and Methods
The trial was conducted in 2012 at the University of Helsinki. The white lupin (Lupinus
albus, var. Ludic) and spring wheat (Triticum aestivum, var. Amaretto) seed density was 36
seeds/m2 and 360 seeds/m2, respectively. The growth was fertilized with 60 kg N/ha. The bicrop was cut with scissors at two growth stages: 96 days (G1) and 110 days (G2) after
sowing. At G1 wheat was at the beginning of the dough stage and at G2 at the end of the
dough stage. Two mixtures of lupin and wheat were made before ensiling: mixture 1 (M1)1/3
lupin + 2/3 wheat and mixture 2 (M2) 2/3 lupin + 1/3 wheat in fresh weight. The mixtures
were treated with three different silage additives: 1) formic acid (FA) 4 l (100%)/tn; 2)
mixture of sodium nitrite (0.75 kg/t) and hexamethylentetramine (hexamine) (0.5 kg/t)
(NaHe); 3) Lactobacillus plantarum 1x106cfu/g forage (LAB). The control (CON) was
without any treatment. The forages were ensiled in 1.5 l glass silos with three replicates per
treatment. Silos were opened after 100 (G1) and 101 (G2) days, respectively. The aerobic
stability of the silages was measured over 12 days. Sample temperature changes compared
with the ambient temperature were monitored using data loggers.
The contamination of silages with clostridia was determined by PCR. For this, silage samples
were stored at -20°C. From 2 to 3 grams of silage was weighed for each DNA extraction.
Silage samples were homogenized in 10 ml milk-lactic-acid-glucose medium. Homogenates
were centrifuged using 10 000 g for 15 min at 23°C. Approximately 200 ng of pellet per
sample was collected for DNA extraction. The DNA extraction was conducted using
Macherey-Nagel NucleoSpin® Soil kit (Macherey-Nagel GmbH & Co. KG, Germany).
Species-specific primers were designed for Clostridium perfringens, C. tyrobutyricum, C.
sporogens, C. butyricum. Quantitative real-time PCR was performed by using Rotor-Gene
6000 (Corbett Research, Australia) with Maxima SYBR Green qPCR Master Mix (Thermo
Proceedings of the 5th Nordic Feed Science Conference
31
Forage conservation and feed processing
Scientific, USA). The lengths of PCR products varied from 254 to 285 bp. All silage samples
were analysed in duplicates and average number of Clostridium genome copies were
calculated per gram of silage. The used DNA extraction protocol should extract DNA from
bacterial spores. Thus, the number of genome copies includes both bacteria and spores.
Statistical tests of silage quality parameters were done for G1M1, G1M2, G2M1 and G2M2
silages separately. Normally distributed variables were tested with ANOVA (SAS 9.3 mixed
procedure). Sums of squares for additive effects were separated using orthogonal contrasts: 1)
CON vs additives; 2) LAB vs FA and NaHe; 3) FA vs NaHe. Non-normal distributed data
were tested with the Kruskall-Wallis analysis (SPSS, version 21). Differences between the
treatments were analysed by pairwise testing.
Results and Discussion
The raw material compositions for different mixtures are given in Table 1. All additives
improved silage quality compared to the control. The performance of the additives varied
between the different growth stages and mix ratio of white lupin and wheat (table 2 and 3).
Comparison of the additives revealed their different functionality. All feed samples were
aerobically stable. The raw material of this trial was contaminated with clostridia spores
despite the fact that the samples were collected unsoiled. This can be deduced from the high
amounts of clostridia in the silages. Whole crop raw materials may be more prone to butyric
acid fermentation than grasses or forage legumes. The contamination with clostridia was also
measured in the numbers of PCR-copies in a gram of silages. From four tested species of
clostridia, C. tyrobutyricum was the most abundant in this trial.
Table 1 Raw material composition (g/kg DM, if not otherwise stated)
Growth stage 1
Growth stage 2
White
Spring
lupin
wheat
Mix 1
Mix 2
White
Spring
lupin
wheat
Mix 1
Mix 2
Dry matter, g/kg
163
379
307
235
138
359
285
Ash
70.2
96.2
91.6
84.2
68.7
88.1
85.0
Crude protein
184
71
91
123
182
61
81
NDF
390
520
497
460
436
525
510
WSC1
126
84
91
103
84
35
43
Starch
30
183
152
113
23
255
218
Buffer cap., mekv/kg DM
709
297
370
488
687
296
359
1
Water soluble carbohydrates; in MIX 1 white lupin - spring wheat ratio 1:2 and MIX 2 ratio 2:1 of fresh
weight.
32
Proceedings of the 5th Nordic Feed Science Conference
212
79.7
114
486
56
154
466
Forage conservation and feed processing
Table 2 The effect of additive on silage quality and clostridia (expressed as log10 copies/g) (g/kg dry matter, if
not otherwise stated). First growth stage
Silage additive1
CON
MIX 1
Dry matter, g/kg
pH
290
4.53
WSC2
29.2
Lactic acid
32.5
FA
311
4.28
109
NaHe
LAB
313
309
5.01
113
2.9
17.2
4.59
10.48
SEM
0.020
0.677
0.801
0.028
<0.001
<0.001
<0.001
23.9
5.70
<0.001
<0.001
0.622
53.0
1.92
0.006
<0.001
<0.001
4.75
0.417
<0.001
<0.001
<0.001
3.75
6.3
Statistical significance1
CON
LAB vs
FA
vs
FA and
vs
additives
NaHe
NaHe
Acetic acid
3.24
Butyric acid
37.69a
6.01ab
0.00b
0.00b
1.297
Ethanol
25.23
3.78
2.32
6.37
0.554
<0.001
0.001
0.100
non-normally distributed
Ammonia N, g/kg N
188
66
107
62
5.2
<0.001
0.005
<0.001
Ammonia N, g/kg N3
188
66
25
62
5.0
<0.001
0.025
<0.001
Clostridia total, lg c/g4
5.61
4.92
2.48
2.07
0.665
0.013
0.081
0.032
C.tyrobutyr., lg c/g4
5.57
4.90
2.50
2.00
0.669
0.014
0.072
0.035
4.1
0.016
0.238
0.556
0.094
0.005
0.002
0.002
0.017
0.006
0.335
7.54
0.974
0.002
0.244
MIX 2
Dry matter, g/kg
pH
226
4.60
237
4.06
2
WSC
14.2
87.2
Lactic acid
45.9
24.9
240
4.67
112
38.4
245
3.83
21.3
75.3
17.1
Acetic acid
7.23
6.79
11.71
7.27
0.718
0.140
0.054
0.001
Butyric acid
43.09
9.32
0.54
0.00
3.201
<0.001
0.244
0.089
Ethanol
28.3
5.92
5.42
0.001
<0.001
0.004
0.815
12.9
Ammonia N, g/kg N
241
64
127
60
10.3
<0.001
0.023
0.003
Ammonia N, g/kg N3
241
64
37
60
10,3
<0.001
0.480
0.104
Clostidia total, lg c/g4
5.66
4.68
3.67
1.93
0.974
0.082
0.097
0.486
C.tyrobutyr., lg c/g4
5.63
4.67
3.67
1.93
0.970
0.084
0.097
0.487
1
CON=no additive, FA=formic acid, NaHe=hexamethylentetramine and sodium nitrite mixure,
LAB=Lactobacillus plantarum; 2 water-soluble carbohydrates; 3 deducted all nitrogen applied through additive;
4
copies per silage gram (fresh weight). In MIX 1 white lupin and spring wheat ratio 1:2 and in MIX 2 ratio 2:1
of fresh weight.
Proceedings of the 5th Nordic Feed Science Conference
33
Forage conservation and feed processing
Table 3 The effect of additive on silage quality and clostridia (expressed as log10 copies/g) (g/kg dry matter, if
not otherwise stated). Second growth stage
Silage additives1
CON
FA
Statistical significance
NaHe
LAB
CON
vs
additives
LAB vs
FA and
NaHe
5.3
0.583
0.090
0.434
0,047
0.001
<0.001
<0.001
SEM
MIX 1
Dry matter, g/kg
315
pH
4.05
311
4.69
317
4.20
327
4.08
FA
vs
NaHe
2
WSC
12.7
33.6
39.1
15.0
2.26
<0.001
<0.001
0.125
Lactic acid
41.9
3.8
32.7
34.7
2.37
<0.001
<0.001
<0.001
Acetic acid
6.37
Butyric acid
4.44
Ethanol
5.80
Ammonia N, g/kg N
Ammonia N, g/kg N
129
3
129
Clostridia total, lg c/g4
4
C.tyrobutyr., lg c/g
a
4.85
11.6
3.99
128
128
ab
9.36
4.26
1.103
0.873
0.068
0.020
0.00
5.69
1.526
0.475
0.960
<0.001
1.61
4.35
0.005
0.002
0.003
0.009
7.5
0.298
0.069
0.728
124
36
b
107
107
ab
7.3
non-normally distributed
5.48
5.90
2.46
6.41
0.773
0.548
0.046
0.014
5.50
3.67
1.87
6.43
0.757
0.122
0.004
0.131
9.5
0.200
0.175
0.857
0.146
0.501
0.693
0.279
MIX 2
Dry matter, g/kg
218
pH
3.93
226
4.20
229
3.96
245
4.00
2
WSC
15.3
65.9
30.5
13.4
8.82
Lactic acid
70.4
5.1
57.8
45.8
8.41
0.008
0.201
0.002
Acetic acid
10.3
8.3
11.7
13.4
2.40
0.770
0.286
0.347
Butyric acid
5.16
Ethanol
11.3
Ammonia N, g/kg N
Ammonia N, g/kg N3
non-normally distributed
8.70
0.43
6.51
3.992
0.991
0.700
0.181
3.41
3.57
7.55
0.869
<0.001
0.005
0.901
155
112
138
130
21.8
non-normally distributed
155
112
44
130
21.7
non-normally distributed
4
Clostridia total, lg c/g
4
C.tyrobutyr., lg c/g
5.60
4.02
2.59
6.53
0.401
0.030
0.000
0.036
5.67
3.73
2.57
6.53
0.341
0.011
<0.0001
0.042
1
CON=no additive, FA=formic acid, NaHe=hexamethylentetramine and sodium nitrite mixure,
LAB=Lactobacillus plantarum; 2 water-soluble carbohydrates; 3 deducted all nitrogen applied through additive;
4
copies per silage gram (fresh weight). In MIX 1 white lupin and spring wheat ratio 1:2 and in MIX 2 ratio 2:1
of fresh weight.
Control vs additives
The values of the control silage quality parameters were clearly worse than those of silages
treated with an additive (P<0.001). The pH value close to 4 of the control silages of G2 did
not prevent high amounts of butyric acid and ammonia but the average of quality parameters
did not differ from the treated silages (P>0.05), because also FA and LAB samples were of
34
Proceedings of the 5th Nordic Feed Science Conference
Forage conservation and feed processing
poor quality. The higher amount of clostridia PCR-copies in the control (P<0.05) compared
with additive treated silages underlined the poor quality of the control silages.
Lactic acid bacteria vs chemical additives
The results of the LAB treatment were good on G1 but poor on G2. Apart from G2M2mixtures the pH value of the LAB silages was lower (P<0.01) than that of the silages treated
with chemical additives. The PCR-copies of clostridia in LAB silages on G2 were
significantly higher (P<0.001) than in silages treated with chemical additives.
Formic acid vs sodium nitrite and hexamine
Both chemical additives restricted fermentation and therefore no difference in WSC content
was observed. The amount of lactic acid was higher in the NaHe silages except in G1M2mixture. The average content of butyric acid was clearly higher in FA samples. However, a
significant difference was observed only in G2M1-mixtures (P<0.001) as a consequence of
high variation of butyric acid content in the replicates of all other mixtures. Ammonia content
of NaHe-silages were higher compared with MH-feeds on the first growth stage (P<0.05).
When adding the NaHe-additive it was assumed that the added nitrogen compounds were
completely broke down to ammonia. G1M1 feeds of NaHe showed lower ammonia values
than the FA-feed (P<0.05) when the total nitrogen amount was corrected by the ammonia of
the added additive. In other feeds the differences were not significant.
The quality of FA silages was poor. Silage pH-value increased during storage time, so that
the highest pH was measured on growth stage 2 of the mixture 1silage (pH 4.69). Formic acid
antimicrobial effect is based on the non-dissociated acid molecule. The pka-value of formic
acid is 3.78. At this pH-level 50 % of the formic acid molecules are dissociated. The sum of
PCR clostridia copies in FA silages was higher in G1 mixture 1 and in G2 both mixtures than
in NaHe silages (P<0.05). In this research, the bactericidal effect of formic acid was not
sufficient due to the high pH of the silages, thus, the ensiling result was likely based on
acidification which, however, was not sufficient. The best silage quality was achieved by
NaHe treatment. NaHe feed quality was good even when feed pH was quite high. The
chemical preservatives of NaHe affect directly the growth of undesired microbes. Thus, the
quality of NaHe-silages was not connected to low pH in the same way as the quality of
silages treated with other additives.
Conclusions
Excluding NaHe treatment, inhibition of butyric acid fermentation was not successful due to
the difficult ensiling properties of the raw material. Despite the better ensiling characteristics
of wheat, compared with white lupin, more wheat in the mixture did not improve ensilability
enough for all the additives to be effective. Lactic acid bacteria based additives worked well
only on the early growth stage when the WSC content was high. Water-soluble carbohydrates
and lactic acid fermentation allowed silage pH to drop down sufficiently to prevent butyric
acid fermentation .The direct acidic effect of formic acid was not sufficient to prevent butyric
acid fermentation. On the other hand, the dosage level prevented lactic acid bacteria activity
and formation of lactic acid, which would have made it possible to maintain a low pH. The
action of a mixture, based on hexamine and sodium nitrite, impaired directly the activity of
undesirable microbes, resulting in silages with almost no butyric acid. The additive prevented
also the activity of lactic acid bacteria and lactic acid production in some samples, resulting
Proceedings of the 5th Nordic Feed Science Conference
35
Forage conservation and feed processing
in silages with relative high pH values. However, pH was not a suitable criterion here for
assessing the efficacy of hexamine and sodium nitrite treated feeds in this research.
References
Azo, W.M., Lane, G.P.F., Davies, W.P. & Cannon, N.D., 2012. Bi-cropping white lupins
(Lupinus albus L.) with cereals for wholecrop forage in organic farming: The effect of seed
rate and harvest dates on crop yield and quality. Biol. Agric. Hortic. 28, 86-100.
Bagge, E., Sternberg Lewerin, S. & Johansson, K-E., 2009. Detection and identification by
PCR of Clostridium chauvoei in clinical isolates, bovine faeces and substrates from biogas
plant. Acta Vet. Scand. 51, 8.
Fraser, M.D., Fychan, R. & Jones, R., 2005. The effect of harvest date and inoculation on the
yield and fermentation characteristics of two varieties of white lupin (Lupinus albus) when
ensiled as a whole-crop. Anim Feed Sci.Technol. 119 307–322.
Hauggaard-Nielsen et al. 2008 Grain legume–cereal intercropping: The practical application
of diversity, competition and facilitation in arable and organic cropping systems. Renew.
Agric.Food Syst. 23, 3–12.
McDonald, P., Henderson, A R. & Heron, S. J. E., 1991. The Biochemistry of Silage. 2nd
edition. Chalcombe Publications, Marlow, UK. 340 p.
Takeuchi, S., Hashizume, N., Kinoshita, T., Kaidoh, T. & Tamura, Y., 1997. Detection of
Clostridium septicum hemolysin gene by polymerase chain reaction. J. Vet. Medic. Sci. 59,
853–855.
Weimer, P.J. & Stevenson, D.M., 2012. Isolation, characterization, and quantification of
Clostridium kluyveri from the bovine rumen. Appl. Microbiol. Biotech. 94, 461–466.
36
Proceedings of the 5th Nordic Feed Science Conference
Forage conservation and feed processing
Protein value in organic grass-clover silages from spring and summer growth
A.K. Bakken1, M. Vaga3, M. Hetta3, Å.T. Randby4 & H. Steinshamn2
1,2
Norwegian Institute for Agricultural and Environmental Research (Bioforsk), 1N-7512
Stjørdal and 2N-6630 Tingvoll, Norway. 3Swedish University of Agricultural Sciences,
Department of Agricultural Research for Northern Sweden, Box 4097, S-904 03 Umeå,
Sweden. 4Norwegian University of Life Sciences, Box 5003, N-1432 Ås, Norway.
Correspondence: [email protected]
Introduction
In organically grown grass-clover swards, there is usually considerable differences in the
content of clover between spring and summer growths. Forage from the first cut is dominated
by grasses and has a lower content of crude protein (CP) and an easily digestible
carbohydrate fraction (Steinshamn and Thuen, 2008). The yields from regrowths contain on
the other hand, more legumes, more indigestible cell walls and more CP. At feeding, these
qualitatively different forages, might be utilized complementary according to the animal
demands. To realize their potential alone or together and explore their complementarity, their
respective values as protein and energy sources have to be determined. Knowing that energy
preservation and prevention of protein degradation during wilting and ensiling are critical for
the final feed value, we have investigated gains and losses obtained by restricted versus
extensive fermentation of the crops. The initial results presented here express feed value
according to NorFor (Volden, 2011).
Materials and Methods
The plant material was harvested from a mixed crop of Phleum pratense, Festuca pratensis,
Lolium perenne, L. boucheanum and Trifolium pratense. The spring growth was harvested at
late stem elongation of the dominating grass species, P. pratense, and the summer growth
614d° afterwards (base temperature 0°C). The content of T. pratense was 30% of dry matter
in spring and 76% in summer growth.
The herbage was wilted indoors for 24 hours before ensiling with different types of additives
(4 mL/kg FM) in evacuated and sealed polyethylene bags. The additives were 1) water
(Control treatment (C)), 2) formic acid (FA, 850 g/kg), and 3) lactic acid bacteria (LAB)
(Kofasil® Lac, Addcon Europe). The dosage of LAB corresponded to 105 cfu per g FM.
Dried (60°C) samples of herbage and silages were analysed for ash, CP, buffer soluble CP
(sCP) (Hedquist and Udén, 2006), neutral detergent fibre (NDF) (Mertens et al., 2002) and
water soluble carbohydrates (WSC) (Larsson and Bengtsson, 1983). Freshly frozen silage
samples were analysed for pH, and content of organic acids and ethanol (Ericsson and André,
2010) and NH4-N. The oven DM contents of the silages were corrected for volatile losses
according to NorFor. Concentrations of indigestible NDF (iNDF) were determined by a 288 h
in situ incubation (Huhtanen et al., 1994). Organic matter digestibility (OMD) was calculated
from iNDF and NDF concentrations (Huhtanen et al., 2013). Net energy lactation (NEL20),
metabolizable protein (calculated as amino acids absorbed in the small intestine (AAT20)) and
protein balance in the rumen (PBV) were calculated according to NorFor at daily intake of 20
kg DM (Volden, 2011). The concentration of constituents in herbage were modelled using the
procedure MIXED in SAS (SAS Institute inc., 1999) with growth (spring or summer) and
wilting as fixed factors and replicate (1-3) as random. For silages, the model included growth
and silage additive as fixed factors and replicate (1-3) as random.
Proceedings of the 5th Nordic Feed Science Conference
37
Forage conservation and feed processing
Results and Discussion
The crop harvested in spring growth, was more digestible and contained less CP and more
WSC than the summer growth dominated by mature red clover (Table 1). All silages were
well fermented as evaluated from pH and the concentration of NH3-N. No single sample
contained more than 72 g NH3-N/kg N, and none had pH higher than 4.50. According to the
content of organic acids, addition of LAB caused the most extensive fermentation (140 g/kg
DM) and FA the least (57 g/kg DM). Protein solubility increased already during wilting in
spring growth, but not in summer growth (Table 1). The CP fraction was further solubilized
during fermentation (Table 2), and the soluble proportion was higher in silages from spring
growth than in silages from summer growth (P<0.001), and higher in extensively fermented
(C, LAB) compared to restrictedly fermented (FA) silages (P<0.001). The restriction of
silage fermentation caused by application of formic acid contributed positively to energy
preservation and protein feed value (NEL20, AAT20), compared to treatments causing more
extensive fermentation (Table 2).
Table 1 Organic matter digestibility (OMD) and content of dry matter (DM), ash, crude protein (CP), soluble
CP (sCP), water soluble carbohydrates (WSC), neutral detergent fiber (NDF) and indigestible NDF (INDF) in a
fresh and wilted grass-clover crop harvested in spring and summer growth. N=3
Herbage,
DM
Ash
CP
sCP
WSC
NDF
iNDF
OMD
growth (G) and wilt (W)
g/kg
g/kg DM
Spring growth, fresh
163
70
101
32
264
391
59
769
Spring growth, wilted
235
76
114
39
221
396
67
759
Summer growth, fresh
139
99
133
40
112
361
113
707
Summer growth, wilted
232
102
138
40
87
383
121
695
SEM
5.0
1.5
2.8
0.9
5.3
7.3
3.0
4.0
Sign. effect of (P≤0.05)
G, W
G, W
G, W
G, W,
G, W
G
G, W
G, W
G×W
Conclusions
Fresh and wilted herbage from a summer growth dominated by red clover contributed more,
and more stable CP than spring growth dominated by grasses did. Preserved as silage, it still
supplied less metabolizable protein (AAT20) because of its lower digestibility. Whether it has
a potential as a complementary forage to silages from spring growth with a lower CP content
needs to be evaluated in vivo. Further investigations which are presently conducted by the
authors, will reveal if the quality and protein value of summer growth silages may be
improved by more frequent or appropriate timing of harvests.
Acknowledgements
The study was financed by the Norwegian Agricultural Agreement Research Fund, Tine BA,
the County governors and councils in Sør- and Nord-Trøndelag, and the Norwegian
Agricultural Extension Service.
38
Proceedings of the 5th Nordic Feed Science Conference
Forage conservation and feed processing
Table 2 Organic matter digestibility (OMD) and content of dry matter (DM), ash, crude protein (CP), soluble
CP (sCP),) metabolizable protein (MP), protein balance in the rumen (PBV) and net energy for lactation
(NEL20) in wilted silages made from spring and summer growth of a grass-clover crop and according to type of
additive.
Silage,
DM
Ash
CP
sCP
OMD
AAT20
PBV
NEL20
growth and additive
g/kg
g/kg DM
Spring growth (n=9)
246
84
124
72
760
80
7
5.9
Summer growth (n=9
237
111
155
70
699
67
56
5.2
SEM
4.9
1.0
3.5
1.9
3.0
0.4
3.4
0.03
Significance, P≤
NS
0.01
0.01
NS
0.001
0.001
0.01
0.001
Control (n=6)
243
99
142
77
731
69
41
5.5
Formic acid (n=6)
240
97
137
62
729
83
13
5.7
Lactic acid bacteria (n=6)
242
98
140
76
729
69
41
5.6
SEM
4.5
0.9
3.4
2.0
3.1
0.4
3.1
0.04
Significance, P≤
NS
0.05
NS
0.001
NS
0.001
0.001
0.05
MJ/kg DM
References
Ericson, B., André, J. & 2010. HPLC-applications for agricultural and animal science.
Proceedings from the 1st Nordic Feed Science Conference, Uppsala, Sweden. Report 274,
Swedish University of Agricultural Sciences, Dept. of Animal Nutrition and Management.
pp. 23-26.
Hedqvist, H. & Udén, P. 2006. Measurement of soluble protein degradation in the rumen.
Animal Feed Science and Technology 126, 1-21.
Huhtanen, P., Kaustell, K. & Jaakkola, S., 1994. The use of internal markers to predict total
digestibility and duodenal flow of nutrients in cattle given six different diets. Animal Feed
Science and Technology 48, 211-227.
Huhtanen, P., Jaakkola, S. & Nousiainen, J., 2013. An overview of silage research in Finland:
from ensiling innovation to advances in dairy cow feeding. Agricultural and Food science 22,
35-56.
Larsson, K. & Bengtsson, S., 1983. Bestämning av lätt tillgängliga kolhydrater i växtmaterial.
Metodbeskrivning 22. Uppsala: Statens Lantbrukskemiska Laboratorium, Sweden.
Mertens, D.R., Allen, M., Carmany, J., Clegg, J., Davidowicz, A., Drouches, M., Frank, K.,
Gambin, D., Garkie, M., Gildemeister, B., Jeffress, D., Jeon, C.S., Jones, D., Kaplan, D.,
Kim, G.N., Kobata, S., Main, D., Moua, X., Paul, B., Robertson, J., Taysom, D., Thiex, N.,
Williams, J. & Wolf, M., 2002. Gravimetric determination of amylase-treated neutral
detergent fiber in feeds with refluxing in beakers or crucibles: Collaborative Study. Journal of
AOAC International 85, 1217-1240.
SAS institute Inc., 1999. SAS/STAT® User’s guide, Version 8, Cary, NC. 3884 pp.
Proceedings of the 5th Nordic Feed Science Conference
39
Forage conservation and feed processing
Steinshamn, H. & Thuen, E., 2008. White or red clover-grass silage in organic dairy milk
production: Grassland productivity and milk production responses with different levels of
concentrate. Livestock Science 119, 202–215.
Volden, H., 2011. Feed calculations in NorFor, in: Volden, H. (Ed), NorFor – The Nordic
feed evaluation system – EAAP 130, Wagenigen Academic Publishers, The Netherlands, pp.
55-58.
40
Proceedings of the 5th Nordic Feed Science Conference
Forage conservation and feed processing
Dynamics of gas formation during ensilage
M. Knicky1, H-G. Wiberg2, F. Eide3 & B. Gertzell3
1
Department of Animal Nutrition & Management, Kungsängen Research Centre, Swedish
University of Agricultural Sciences, S-753 23 Uppsala, Sweden; 2Dept. Analyses &
Environmental service, Coor Service Management, S-251 09 Helsingborg, Sweden; 3AB
Hanson & Möhring, P.O. Box 222, S-301 06 Halmstad, Sweden
Correspondence: [email protected]
Introduction
The formation of a whole spectrum of gasses occurs during ensiling process. The formation
of gasses is undesirable, because it is often a sign of undesirable processes in silages, and
causes concern about the impact on the global environment. Formation of CO2 is the most
abundant during ensiling and its volume can reach up to 80% of total gasses produced during
the first 60 ensiling days (Peterson et al., 1958). Processes where CO2 occurs as a by-product
are less effective in transformation of substrate to main fermentation products which results
in higher ensiling losses (McDonald et al., 1991). Therefore, the formation CO2 can be
considered as a measure of ensiling efficiency and ensiling losses. Besides CO2, the
formation of toxic N oxides also occurs during ensiling. Research concerning nitrogenous gas
formation during ensiling is sporadic and often incomplete (Spoelstra, 1985). Since silage
additive addition affects the fermentation pattern, the aim of the study was to monitor the
formation of various gasses from silages treated with different silage additives.
Materials and Methods
A grass ley (70% timothy) was harvested on August 24th 2013, nearby Helsingborg, southern
Sweden. The forage crop was directly harvested using a precision chopper (Claas Jaguar
690). Standard analyses to determine composition of the fresh forage (FF) such as dry matter
(DM), buffering capacity, ash, water soluble carbohydrates (WSC), pH, neutral detergent
fibre (NDF), metabolisable energy (ME), crude protein (CP), and hygienic quality
(enterobacteria, clostridia spores) were performed. The forage was divided in three fractions.
One was left untreated while the rest was treated with either of two silage additives; one with
a bacterial inoculant (E. faecium, L. plantarum, L. buchneri) at the rate of 250000 cfu/g FF,
and the second one with Safesil at the rate of 3 L/ton FF. Forages were ensiled in steel
laboratory silos (vol. of 25 L). Each treatment consisted of 6 replicates. Bottom and lids of
silos were equipped with stoppers with tubes allowing collection of gas and silage liquids.
Two forms of gas collection were applied. First, the escaping gasses in three silos of each
treatment were collected into Tedlers bags (Supelco), which were regularly changed. Gasses
collected in bags were analysed for N2, H2, O2, CO, CH4 by gas chromatographaphy.
Separation was done on a packed column (40/60 mesh, 4 m, OD 1/8) in a Perkin Elmer
Clarus 580 gas chromatograph using TCD detector. The second set of three silos of each
treatment was connected to distilled water baths (regularly changed) in which CO2, NO, and
NO2 were absorbed. The gases absorbed in water were analysed using ion chromatography by
use of a UV detector according to ISO 10304-1. Gas collections were done during14 days
(bacterial treatment for 30 days). At the end of storage (120 days), silages samples were
extracted and standard analyses (DM, volatile fatty acids, lactic acid, ethanol, pH, NDF, ME,
crude protein, lactic acid bacteria, clostridia spores, aerobic stability) were performed to
determine silage quality.
Proceedings of the 5th Nordic Feed Science Conference
41
Forage conservation and feed processing
Results and Discussion
The chemical and microbiological composition of the forage, prior to ensiling, is presented in
Table 1. Chemical composition of fresh forage represented a common composition found in
third cut grass crops in Sweden. Calculated fermentation coefficient of 39 characterizes the
forage as slightly above the limit for a difficult crop to ensile (Weissbach, 1974). The
analyses of microbiological contamination revealed high counts of enterobacteria and
particularly of clostridia spores.
Table 1 Chemical and microbiological compositions of fresh forage (n=2)
Analyses
Unit
DM
%
23.8
Ash
% of DM
8.6
CP
% of DM
13.8
WSC
% of DM
11.3
NDF
% of DM
55.9
ME
MJ/kg DM
10.9
Buffering capacity
g LA/100 g DM
6.0
Enterobacteria
log cfu/g FM
3.4
Clostridia spores
log cfu/g FM
4.7
pH of forage mass
6.2
Fermentation coefficient
39
DM-dry matter; FM-fresh matter; CP-crude protein; WSC-water-soluble carbohydrates; NDF-neutral detergent fiber;
ME-metabolisable energy.
Results of chemical analyses of the silages displayed variation in chemical composition (Table 2).
Silages treated with bacterial inoculant had significantly higher pH, propionic acid, acetic acid,
and ethanol contents but lower concentration of lactic acid than other silage treatments
Table 2 Chemical composition of silages at the end of storage (n=3)
DM
pH
%
NH3-N
Lactic acid
% TN
% DM
Acetic acid
Propionic acid
Ethanol
Control
24.3
4.0
6.6
10.1
1.7
0.03
0.5
Inoculant
24.1
4.4
7.1
3.9
5.5
0.45
0.8
Safesil
25.1
4.0
6.2
8.7
1.8
<0.04
0.3
LSD0.05
0.02
1.24
0.38
0.20
0.02
0.05
P-value
0.001
0.3
0.001
0.001
0.001
0.001
It is assumed that the effects were caused by L. buchneri in the silage innoculant which is known
to form acetic acid at the expence of lactic acid (Reich & Kung, 2010). As a concequence of the
high levels of acetic acid, silage treated with bacterial innoculant were found to be significanly
more aerobicaly stable than the untreated control silage (Table 3). This study confirmed early
findings (Knicky & Spörndly, 2009, 2011) that Safesil is efficient to secure a proper ensiling
process and in improving the aerobic stability of silage.
42
Proceedings of the 5th Nordic Feed Science Conference
Forage conservation and feed processing
Table 3 Microbiological composition and aerobic stability of silages at the end of storage (n=3).
Time (hours) until temp.
aerated silages increased 3°C
Max. temp
(°C)
Max. temp. increase
(°C)
pH after
stab.
Yeasts (lg
cfu/g)
72
31.8
13.9
6.1
3.6
4.5
<1.7
Control
Inoculant
166
19.6
1
Safesil
166
19.3
0.3
4.1
<1.7
LSD0.05
16.8
1.37
0.32
P-value
0.001
0.02
0.001
Gas analyses revealed significantly higher formation of total gasses in control silage and bacterial
innoculant treated silage than in the Safesil treated silage (Figure 1). This could be explained by
differences in fermentation intensity and variation in bacterial composition among the silage
treatments. A high gas formation in bacterially inoculated silages is assumed to be a consequence
of addition of bacterial microflora which intensified the fermentation process. On the other hand,
Safesil possesses a rather selective inhibitory property which restricts particularly undesirable
fermentation processes. This probably caused a lower gas formation in Safesil treated silages. The
proportion of CO2 of total gass was 0.56 for Safesil, 0.65 for bacterial inoculant, and 0.68 for
control silage and similar to Peterson et al. (1958).
All silages displayed a similar pattern in development of gas over time. The highest formation of
gasses was observed approximatelly between 11-29 hours of the ensiling period. This peek of gas
formation corresponds to the period of the most intensive fermentation, according to Pahlow et al.
(2003). Another increase in gas formation was observed in bacterialy innoculated silages after ca.
300 hours of ensiling. This increase in CO2 formation was observed only in these silages (Figure
2). This phenomenon is regarded to be the consequence of L. buchneri activity, which has the
ability to convert lactic acid into acetic acid under unearobic conditions (Oude Elfering et al.,
2001) and where CO2 is formed as a bi-product of this conversion.
Due to lack of formation of ensiling liquid, it was impossible to follow degradation of nitrate
during ensiling process. Neverheless, the formation of NO (Figure 3) and NO2 (Figure 4) seemes
to follow the degradation of nitrate and nitrite during fermentation, as described by Spoelstra
(1985) and demonstrated by Knicky & Lingvall (2005) and Knicky & Spörndly (2009). The
study of Peterson et al. (1958) showed a similar pattern of NO formation as in the present study.
Surpisingly, the formation of all NOx gasses revealed no statistical differences between Safesil
and the other silages. A higher formation of NO was expected due to the presence of Na-nitrite in
Safesil. The lack of N2O measurement as well as monitoring of ammonia formation during
fermentation process makes it difficult to explain the lack of variation in NOx formation.
Conclusions
All silages were well fermented; however, bacterially inoculated silages contained increased
concentrations of acetic and propionic acid and ethanol. Both additive treated silages possessed
improved aerobic stability. Control and bacterially inoculated silages produced more gas than
Safesil treated silages, mainly due to an increased proportion of CO2. The formation of NOx gases
displayed no significant differences among other treatments.
Proceedings of the 5th Nordic Feed Science Conference
43
Forage conservation and feed processing
Total gasses
6.00
5.00
ml/l
4.00
3.00
2.00
1.00
0.00
Hours
Untreated
Inoculants
Safesil
Figure 1 The sum of all gasses formed in silages during measurement. (n=3)
CO2
7000
6000
mg/l
5000
4000
3000
2000
1000
0
Hours
Untreated
Inoculant
Safesil
Figure 2 The formation of CO2 in silages during measurement. (n=3)
44
Proceedings of the 5th Nordic Feed Science Conference
Forage conservation and feed processing
NO
2500
mg/l
2000
1500
1000
500
0
Hours
Untreated
Inoculant
Safesil
Figure 3 The formation of NO in silages during measurement. (n=3)
NO2
2500
mg/l
2000
1500
1000
500
0
Hours
Untreated
Inoculant
Safesil
Figure 4 The formation of NO2 in silages during measurement. (n=3)
Proceedings of the 5th Nordic Feed Science Conference
45
Forage conservation and feed processing
References
Knicky, M & Lingvall P. 2004. Ensiling of high wilted grass-clover mixture by use of different
additives to improve the hygienic quality. Acta Agr.Scand., Section A, 54, 197-205.
Knicky, M. & Spörndly, R., 2009. Sodium benzoate, potassium sorbate and sodium nitrite as
silage additives. J. Sci. Food Agric. 89, 2659-2667.
Knicky, M. & Spörndly, R., 2011. The ensiling capability of a mixture of sodium benzoate,
potassium sorbate, and sodium nitrite. J. Dairy Sci., 94, 824-831.
McDonald, P., Henderson, A. R., & Heron, S. J. E., 1991. The Biochemistry of Silage.
Chalcombe Publications, Marlow, Bucks, UK.
Oude Elfering, S.W.J.H., Krooneman, J., Gottschal, J.C., Spoelstra, S.F., Faber, F. & Driehuis, F.,
2001. Anaerobic Conversion of Lactic Acid to Acetic Acid and 1,2-Propanediol by Lactobacillus
buchneri. Appl. Environ. Microbiol. 67, 125-132.
Pahlow, G., Muck, R.E., Driehuis, F. & Oude Elferink, S.J.W.H. 2003. Microbiology of ensiling.
Ed. Buxton, D.R., Muck, R.E. & Harrison, J.H. 2003. Silage science and technology, Madison,
Wisconsin, USA.
Reich, L.J. & Kung, L. Jr., 2010. Effects of combining Lactobacillus buchneri 40788 with various
lactic acid bacteria on the fermentation and aerobic stability of corn silage. J. Anim. Feed Sci.
Technol., 159, 105-109.
Spoelstra, S.F., 1985. Nitrate in silage (review). Grass Forage Sci. 40, 1-11.
Weissbach, F., Schmid, L. & Hein, E., 1974. Method of anticipation of the run of fermentation in
silage making based on chemical composition of green fodder. Pages 663-673 in Proc 12th Intern
Grassland Congress, Moscow, Russia, 11-20 June.
Peterson, W.H., Burris, R.H., Sant, R. & Little, H., 1958. Production of toxic gas. (Nitrogen
oxides) in Silage Making. Agricultural Food Chemistry, 6, 121-126.
46
Proceedings of the 5th Nordic Feed Science Conference
Forage conservation and feed processing
Mycotoxins in haylage
J. Schenck1,2, C. Müller1 & R. Spörndly1
1
Swedish University of Agricultural Sciences (SLU), Department of Animal Nutrition &
Management, Feed Science Division, Kungsängen Research Centre, 753 23 Uppsala.
2
Department of Forest Mycology and Plant Pathology, Swedish University of Agricultural
Sciences, BOX 7026, 750 07 Uppsala, Sweden
Correspondence: [email protected]
Introduction
Silage with high dry matter content, also referred to as haylage, is a common forage in feed
rations for horses in Sweden (Enhäll et al., 2012). As water activity is low in haylage, lactic
acid fermentation is restricted resulting in high content of residual water soluble
carbohydrates and no or a very small pH-decrease as opposed to silage (Finner, 1966; Müller,
2005). If haylage bales are not sufficiently air-tight, this environment may favour mould
growth in the forage. Mould growth may increase the risk for mycotoxin presence in the feed
as well, but there is scarce information about formation of mycotoxins in haylage and in
horse feeds in general (Liesener et al., 2010). Moulds do not regularly produce mycotoxins,
but may occur as secondary metabolites under certain environmental conditions (Samson et
al., 2002). The aim of this study was therefore to examine mycotoxin presence in randomly
selected haylages in Sweden and Norway, and to investigate if any correlations existed
between mycotoxin and mould presence.
Materials and Methods
Haylage samples from in total 100 farms in Sweden and Norway were analysed for chemical
composition, mould growth and mycotoxins. Sampling of haylage was performed at 77
different Swedish farms during two years (2010 and 2011), and at 23 Norwegian farms
during 2011. The farms were evenly distributed over the countries. Samples for mould
cultivation were taken in three ways as described by Schenck et al. (2013). In short, they
were: i) direct plating of samples taken from visible fungal growth on bale surface, ii) direct
plating of plant material from drilled core samples and iii) dilution plating where core
samples were mixed with peptone water and the dilutions cultured. Two substrates (malt
extract agar (MEA) and Dichloran 18% glycerol agar plates (DG-18) (Merck, KGaA,
Darmstadt, Germany)) and at two incubation temperatures (25 and 37 °C) were used for
culturing. Samples were plated within 48 hours after sampling, and were incubated for ten
days. For mycotoxin analysis, core samples were used. These samples were ground after
freeze-drying and analysed by LC-MS/MS as described by Rasmussen et al. (2010) at
National Food Institute, Division of Food Chemistry, Technical University of Denmark.
Eleven mycotoxins, including patulin, nivalenol (NIV), deoxynivalenol (DON), 3acetyldeoxynivalenol (15-ACDON)), gliotoxin, alternariol, HT-2 toxin, T-2 toxin,
zearalenone (ZEA), beauvericin (BEAU) and enniatin B (ENN B), were analyzed. Lowest
detection limits were: 371 μg patulin/kg, 122 μg NIV/kg, 20 μg DON/kg, 35 μg 15ACDON/kg, 41 μg gliotoxin /kg, 10 μg alternariol/kg, 5 μg HT-2 toxin/kg, 8 μg T-2 toxin/kg, 5
μg ZEA/kg, 10 μg BEAU/kg and 10 μg ENN B/kg.
Chemical variables were analysed as described by Müller (2005).
Proceedings of the 5th Nordic Feed Science Conference
47
Forage conservation and feed processing
Correlation calculations between presence of mycotoxins and moulds in haylage samples
were performed using PROC CORR statement (SAS, 2014). The probability to find
mycotoxins was also tested with PROC LOGISTIC model (SAS, 2014), where variables such
as presence of mould (any species or Fusarium spp. in particular) and of chemical
composition such as dry matter, crude protein, ash, NDF, pH, ammonia nitrogen, lactic acid,
acetic acid, propionic acid, butyric acid, ethanol and 2,3-butanediol were included. In all
statistical calculations, farm was used as experimental unit and effects were considered as
statistically significant when P<0.05.
Results and Discussion
One or more mycotoxin(s) were found in haylage from fifty farms, while no mycotoxins were
detected in the remaining haylage samples (50) (Figure 1). The most frequently detected
mycotoxin was ENN-B (31 farms) followed by BEAU (16 farms) and DON (12 farms).These
mycotoxins are all known to be produced by Fusarium species. Minimum, maximum and
mean mycotoxin concentrations are reported in Table 1. The level of the second most
common mycotoxin found in the present study (BEAU) was low, compared to a recently
presented survey from Korea, where the average level of concentrate feeds for cattle was 720
µg/kg (Kyung et al 2010). Wheat bran was identified as the concentrate ingredient containing
the highest amount BEAU ranging from 340 to 1100 µg/kg.
60
Number of farms
50
40
30
20
10
0
Mykotoxin, µg/kg
Figure 1 Mycotoxins detected in haylage sampled from 100 farms in Sweden and Norway and number of farms
where each mycotoxin was present.
Mould was detected in haylage samples from 49 farms with method ‘’', 74 farms with method
‘ii’, and 56 farms with method ‘iii’. The most common mould genera were Pencillium spp.
and Artrinium spp. Fusarium spp. were only detected in haylage from five farms.
48
Proceedings of the 5th Nordic Feed Science Conference
Forage conservation and feed processing
Table 1 Mycotoxins detected in haylage samples from 100 Swedish and Norwegian farms (µg/kg). Values only
represent samples where the mycotoxins were detected
Mycotoxin
Minimum
Maximum
Mean
SD
Patulin
Nd
Nd
Nd
Nd
NIV
Nd
Nd
Nd
Nd
DON
69
479
238
134.7
15- ACDON
70
288
179
154.1
Gliotoxin
44
57
51
9.2
Alternariol
11
1452
212
501.7
HT-2
19
78
35
28.8
T-2
8
11
9
1.5
ZEA
8
8
8
-
BEAU
11
988
248
376.8
ENN B
10
283
56
83.9
Nd = value below lower limit of detection (see text for lower detection limits)
There was no correlation between mycotoxin presence and mould occurrence (r=0.03). Thus,
finding visible mould at the bale surface or from core samples will not mean that mycotoxins
are present. An analysis of the magnitude of total mould growth (log CFU/g) established with
the dilution method, and the prevalence of mycotoxins, resulted in an indication that the risk
of finding mycotoxins was higher when mould numbers were higher (r=0,25; P<0,05).
However, higher mould counts were not correlated to higher mycotoxin concentrations
(r=0.03). This is in contrast to the findings of Legzdina and Buerstmayr (2004) who studied
Fusarium head blight and mycotoxin presence in barley, where the percentage of visually
infected barley spikes was positively correlated with the content of DON and 15-ACDON.
It was notable that no correlation was found between occurrence of Fusarium-toxins (NIV,
DON, BEAU and ENN-B) and presence of Fusarium spp.-moulds (Fusarium culmorum,
Fusarium equiseti, Fusarium graminearum and Fusarium poae) in the haylage samples.
However, the small number of haylage samples where Fusarium-moulds were detected may
partly explain this. Also, presence of mycotoxin may be a result of previous and not present
active mould growth, meaning that the product of the mould is detected but not the mould
itself.
The difficulty of using other variables as predictors of mycotoxin presence was further
highlighted when analysing the data with the LOGISTIC procedure, where the statement
SELECTION=FORWARD was used, allowing all chemical variables and mould occurrence
that met the significance level of P<0,05 to enter the model. No variable appeared to
statistically increase or decrease the odds ratio to find mycotoxins in the haylage.
Conclusions
Mycotoxins were detected in 50 % of sampled haylage bales in Sweden and Norway.
Fusarium spp. mycotoxins were the most common mycotoxins present. Neither conventional
mould analysis by culturing nor chemical variables were able to indicate if mycotoxins were
Proceedings of the 5th Nordic Feed Science Conference
49
Forage conservation and feed processing
present. However, in bales where moulds occurred, higher mould counts correlated weakly
with the presence of mycotoxins.
References
Enhäll, J., Nordgren, M. & Kättström, H., 2012. Hästhållning i Sverige 2010. The Swedish
Board of Agriculture, Report 2012:1. Jönköping, Sweden.
Finner, M.F., 1966. Harvesting and handling low-moisture silage. Transact. ASAE 9, 377381.
Legzdina, L. & Buerstmayr, H., 2004. Comparison of infection with Fusarium head blight
and accumulation of mycotoxins in grain of hulless and covered barley. J Cereal Sci 40, 6167.
Liesener, K., Curtui, V., Dietrich, R., Märtlbauer, E. & Usleber, E. 2010. Mycotoxins inn
horse feed. Mycotox. Res. 26, 23-30.
Kyung-Eun, L., Byung Hee, K. & Chan, L., 2010. Occurrence of Fusarium mycotoxin
beauvericin in animal feeds in Korea. Anim. Feed Sci. Technol., 157, 190-194
Müller, C. E., 2005. Fermentation patterns of small-bale silage and haylage produced as a
feed for horses. Grass Forage Sci. 60, 109-118.
Rasmussen, R.R., Storm, I.M.L.D., Rasmussen, P.H., Smedsgaard, J. & Nielsen, K.F., 2010.
Multi-mycotoxin analysis of maize silage by LC-MS/MS. Anal. Bioanal. Chem. 397, 765776.
Samson, R.A., Hoekstra, E.S., Frisvad, J.C., Filtenborg & O., 2002. Introduction to food- and
airborne fungi. 6th ed. Centralbureau voor Schimmelcultures, Utrecht, The Netherlands.
SAS, 2014. SAS 9.3. SAS institute Inc. Cary. NC. USA
50
Proceedings of the 5th Nordic Feed Science Conference
Forage conservation and feed processing
Energy efficient and competitive on farm storage of moist feed grain - preliminary
results
N. Jonsson1, J. Blomqvist2 & M. Olstorpe2
1
JTI – Swedish Institute of Agricultural and Environmental Engineering, P.O. Box 7033, S750 07 Uppsala, Sweden, 2Department of Microbiology, Swedish University of Agricultural
Sciences, S-750 07 Uppsala, Sweden
Correspondence: [email protected]
Introduction
Feed grain is the principal source of energy (40-80%) in animal feeds. Preparation of on-farm
feed grain is profitable if a high quality feed is produced. Fungi are the major factor
restricting the storability of feeds and fungal activity may lead to a decrease in nutritional
value and loss of palatability, and to health hazards for animal consuming the feed and people
handling it. Storage fungi are quite common in feed grain preserved on Swedish farms
(Jonsson & Pettersson, 1999). Depending on preservation method used, 20-60% of the farms
produced feed grain contaminated with storage fungi. Many of the storage fungi are potential
producers of mycotoxins. High-temperature drying is probably the safest preservation method
but involves high investment and energy costs. An alternative is sealed storage of the grain in
silos, which is an environmentally more beneficial method and can be used up to moisture
contents (MC) of 26 %. This method only requires about 2% of the energy consumed in hightemperature drying (Pick et al., 1989). Sealed storage relies on perfect sealing in combination
with the anaerobic conditions caused by respiration of the grain itself and the adhering
microorganisms. However, it is difficult to achieve fully anaerobic conditions since leakage,
the removal of grain and great diurnal changes in gas pressure result in gas being exchanged
between the stored and the surrounding atmosphere (Ekström, 1992; Newman, 1971;
Druvefors et al., 2002). An experience is that grain stored in airtight silos usually deteriorates
when stored until the following summer. A way of prolonging storage time is the use of
propionic acid at levels of 3-8 l/tonne, depending on water content (Ekström, 1992). Another
method for grain with high MC (>30 %), increasingly used, is storage of crimped grain in
silo-bags. Here, the grain is converted into a stable livestock feed by fermentation (Finch et
al., 2002). With this method, growth of fungi and bacteria typically occurs for only a short
period during initial stages of fermentation after which lactic acid production and alcoholic
fermentation pre-dominate (Salovaara, 1998). Pathogenic and spoilage organisms may be
inhibited by production of organic acids, hydrogen peroxide, carbon dioxide and/or
antimicrobial substances, and also by lowering the pH (Ouwehand, 1998). A group of
bacteria, which have been identified in moist stored grain are Enterobacteriaceae (Olstorpe et
al., 2012). Many members of this family are a normal part of the gut flora of humans and
other animals, while others are found in water or soil, or are parasites on a variety of different
animals and plants. This group include EHEC, a type of Escherichia coli, associated with
severe human health problems. Lowering the number of these bacteria in feeds has been
shown to reduce their frequency in the food chain (Brooks et al., 2001). A problem with a
preservation method of high MC grains is that during favourable harvest conditions, the MC
will decrease with up to 5% units per day (Gunnarsson et al., 2012) and only a part of the
stored grain will be protected by fermentation, unless water is added during the crimping.
Drier conditions will favour the development of fungi and one of the most important spoilage
fungi in airtight stored grain is Penicillium roqueforti. This fungus can grow at low partial
pressures of oxygen (>0.14%) (Magan and Lacey, 1984), high levels of carbon dioxide and
Proceedings of the 5th Nordic Feed Science Conference
51
Forage conservation and feed processing
low temperatures and is a potential mycotoxin producer (Häggblom, 1990; Pitt and Hocking,
1999).
In the present study, the hygienic quality of moist crimped barley and wheat (in mixture or
separate) stored in silo-bags on three Swedish farms were evaluated during two successive
years. As in an earlier study, which showed that the microbial flora in grain stored in this
kind of system is highly diverse and not predictable (Olstorpe et al., 2010), the microbial
flora were identified and enumerated. Compared with the earlier study in which the storage
ended in March, the objective was to follow the grain quality from harvest until July the
following year. The evaluation of the storage system will also include technology and
economy.
Materials and Methods
Experimental design and feed components
Three Swedish farms, two in Uppland and one in Hälsingland, took part in the study. All
farmers used a bagging machine combined with a crimper: Murska with rollers (Farm 1 and
3) or Bagmaster with discs (Farm 2), and an auger for packing the crimped grain into the silobag. Both years, the harvest period started in August (Uppsala) and ended in October
(Hälsingland). In most cases, the grain was harvested (wheat or barley, separate or in
mixture), crimped and packed into silo-bags during the same day. In Uppland, both farms
(Farm 1 and 2) added propionic acid to the grain (3-5 l/ton) as an extra precaution. The silobags had a diameter of 2 (Farm 1 and 3) or 2.4 m (Farm 2) and a length of up to about 50-60
m, containing up to 200 ton. To ensure proper fermentation of the grain, the tubes were not
opened until after 3-4 weeks of storage. After opening, a layer of about 0.5 m was removed
daily from the face. Grain samples were collected from each farm at harvest (start), in
October/November, December, March/April, May, June and July by JTI. In connection to the
sampling in November to July, gas-composition (CO2 and O2; GA2000 Gas Analyser
Geotechnical Instruments Ltd, UK) and the temperature (TSI VelociCalc 9565, TSI
Incorporated, MN, USA) of the grain were measured. Information of average daily
temperatures during the storage periods were received from the Swedish Meteorological and
Hydrological Institute.
Analytical methods
Moisture content was determined by drying samples at 103 ◦C for 16 h and water activity of
the grain samples by an aw-meter (AquaLab Series 3TE, Decagon Devices Inc., Pullman,
WA, USA). The pH of the liquid obtained after mixing aliquots of 100 g grain with 100 ml
deionized water and shaking for 10 minutes at room temperature was measured using a
laboratory pH meter (Radiometer, Copenhagen, Denmark).
Microbial quantification
Samples for microbial quantification, etc. were collected from the surface where the grain
was unloaded (3 points). From these samples, aliquots of 20 g of barley were suspended in
180 ml of peptone water (2 g/l) and homogenized with a stomacher as previously described
(Petersson et al., 1999). The samples were serially diluted and spread on selective agar plates
to enumerate and isolate LAB, yeasts, moulds, , Enterobacteriaceae. LAB were quantified by
anaerobic incubation at 30°C for 48 h on de Man Rogosa Sharp (MRS) agar (Oxoid Ltd.,
Basingstoke, UK) supplemented with 100 g/l Delvocid (Gist-brocades B.V., Delft, The
Netherlands). A GasPack system (Becton Dickinson; Sparks, MD, USA) was used during the
52
Proceedings of the 5th Nordic Feed Science Conference
Forage conservation and feed processing
incubation to obtain an anaerobic environment. Yeasts were enumerated on Malt Extract
Agar (MEA) (Oxoid Ltd.) plates, supplemented with 100_g/l chloramphenicol (Sigma–
Aldrich, Stockholm, Sweden). Plates were incubated at 25°C for 2–4 days. Moulds were
quantified on Dichloran-Glycerol (DG18) Agar after incubation at 25°C for 5-7 days.
Enterobacteriaceae were enumerated on Violet Red Bile Glucose (VRBG) Agar (Oxoid Ltd.)
incubated at 37°C for 24 h, by the pour plate method (1 ml sample overlaid with culture
media; Lyberg et al., 2008; Olstorpe et al., 2010). Colonies were counted and mean cfu/g
cereal grain was calculated. Moulds were identified both based on morphological
characteristics and sequencing (Leong et al 2012, Samson et al. 2004).
Yeast and LAB strain conservation
To identify the dominating yeasts and LAB flora in the storage systems, 10 yeasts and 10
LAB colonies were randomly selected from quantification plates at the start and end of the
storage period. Yeast isolates were conserved by mixing 1 ml of an overnight culture of the
yeast isolate with 1 ml glycerol and frozen at -80°C. LAB strains were conserved in a cryo
solution and frozen as described by Olstorpe et al. (2008).
DNA extraction, amplification and identification of microorganisms
Yeast and lactic acid bacteria and mould DMA were extracted separately according to Leong
et al. (2012). For sequencing of the chosen yeast and LAB isolates, the D1/D2 region of the
26S subunit respectively 16S region of the ribosomal DNA were amplified using NL1/NL4primers (yeast) and 16S/16Sr-primers (LAB) (Olstorpe et al. 2008). PCR for mould
sequencing was done according to Leong et al (2012). A volume of 21 µl of RNase free
water, 1 µl of each primer and 2 µl of either extracted DNA (yeast, moulds) or cell
suspension (LAB) were pipetted in a Illustra PuReTaq Ready-to-go PCR bead-tube. PCR was
run according to Olstorpe et al (2008, 2010). The amplified sequences were sent to Macrogen
Europe Inc (Amsterdam, The Netherlands) and the isolates were identified by sequence
comparison against the NCBI´s database (http://blast.ncbi.nlm.nih.gov/Blast).
Results and Discussion
Preliminary results show considerable variations in MC and gas composition in the stored
grain, Table 1. Elevated levels of oxygen could usually be linked to damage of the silage
casing caused by birds or by the unloading tractor. The addition of propionic acid resulted in
lower microbial activity and thus usually resulted in higher oxygen content and lower content
of carbon dioxide.
The addition of propionic acid, the same quantity regardless of MC in the grain, reduced the
occurrence of yeasts, moulds and Enterobacteriaceae, often to below the detection level. The
presence of lactobacilli remained mostly constant. When propionic acid was not used, the
prevalence of lactobacilli usually increased substantially while yeasts and Enterobacteriaceae
decreased slightly, and the presence of moulds decreased to below the detection limit.
Traffic within the farm would need more attention. Often, the transport of feed and fertilizer
crossed each other with an increased risk of contamination of the feed with manure. Farm 3
also lacked a paved open area (concrete/asphalt) for the storage of the silo-bags, which
increase the risk of soil/manure contamination in connection with handling of the feed.
Proceedings of the 5th Nordic Feed Science Conference
53
Forage conservation and feed processing
Table 1 Preliminary results concerning treatments, gas composition and the development of microorganisms in
crimped moist feed grain during storage in silo bags on three Swedish farms
Farm 1 (Uppland) (mixture of wheat and barley)
2011/2012
2012/2013
17-26
24-39
3-4
3-4
6-20
4-21
O2, % 10 m from unloading point
0,6-14
0-19
CO2, % 1m from unloading point
3-17
0-20
13-29
3-53
Yeasts, (log10CFU/g)
<2
<2
Lactic acid bacteria, (log10CFU/g)
3- 6
<2-7
Moulds, (log10CFU/g)
<2
< 2 - 2.3
Enterobacteriaceae, (log10CFU/g)
<1
< 1 – 1.5
17-36
19-22
5
4
O2, % 1m from unloading point
5-21
0-21
O2, % 10 m from unloading point
2-17
0-21
CO2, % 1m from unloading point
0.7-18
0-23
9-18
0.3-18
Yeasts, (log10CFU/g)
<2–6
< 2- 3
Lactic acid bacteria, (log10CFU/g)
<2–6
< 2- 3
Moulds, (log10CFU/g)
<2–4
< 2- 2
Enterobacteriaceae, (log10CFU/g)
<2–5
< 1- 3
MC, % at harvest
Propionic acid, l/ton
O2, % 1m from unloading point
CO2, % 10 m from unloading point
Farm 2 (Uppland) (mixture of wheat and barley)
MC, % at harvest
Propionic acid, l/ton
CO2, % 10 m from unloading point
54
Proceedings of the 5th Nordic Feed Science Conference
Forage conservation and feed processing
Table 1 (Continued)
Farm 3 (Hälsingland) (barley)
2011/2012
MC, % at harvest
2012/2013
17-27
20-34
0
0
O2, % 1m from unloading point
5-17
4-21
O2, % 10 m from unloading point
0-0.4
0-21
CO2, % 1m from unloading point
8-34
0-30
27-54
1-83
Yeasts, (log10CFU/g)
<2–6
<2–6
Lactic acid bacteria, (log10CFU/g)
<2–8
2–7
Moulds, (log10CFU/g)
<2–5
<2
2–7
<1–7
Propionic acid, l/ton
CO2, % 10 m from unloading point
Enterobacteriaceae, (log10CFU/g)
Farm 3 (Hälsingland) (wheat)
MC, % at harvest
29-43
Propionic acid, l/ton
0
O2, % 1m from unloading point
0-21
O2, % 10 m from unloading point
0-17
CO2, % 1m from unloading point
0.4-68
CO2, % 10 m from unloading point
12-98
Yeasts, (log10CFU/g)
<2–5
Lactic acid bacteria, (log10CFU/g)
6–8
Moulds, (log10CFU/g)
<2-3
Enterobacteriaceae, (log10CFU/g)
<1–5
References
Brooks, P.H., Beal, J.D. & Niven, S.J., 2001. Liquid feeding of pigs: potential for reducing
environmental impact and for improving productivity and food safety. Rec. Adv. Anim. Nutr.
Aust. 13, 49–63.
Druvefors U., Jonsson, N., Boysen, M. & Schnürer, J., 2002. Efficacy of the biocontrol yeast
Pichia anomala during long-term storage of moist feed grain under different oxygen and
carbon dioxide regimens. FEMS Yeast Res. Aug;2(3):389-94.
Ekström, N., 1992. Lufttät lagring av fuktig foderspannmål. Meddelande nr 439, JTI,
Uppsala.
Finch, H.J.S., Samuel, A.M., Lane, G.P.F. & Wiseman, A.J.L., 2002. Cereals. In: Lockhart
and Wiseman’s Crop Husbandry including Grasslands. Woodhead Publishing Limited,
Cambridge, England, pp. 259–302.
Gunnarsson, C., de Toro A., Jonsson N. & Lundin G., 2012. Cereal harvesting- strategies
and costs under variable weather conditions. JTI-rapport 403, Lantbruk och industri.
Proceedings of the 5th Nordic Feed Science Conference
55
Forage conservation and feed processing
Jonsson, N. & Pettersson, H., 1999. Evaluation of different cereal grain conservation methods
based on hygienical quality. Swedish Institute of Agricultural and Environmental
Engineering, Uppsala, Sweden, Report 263 (in Swedish).
Lyberg, K., Olstorpe, M., Passoth, V., Schnürer, J. & Lindberg, J.E., 2008. Biochemical and
microbiological properties of a cereal mix fermented with whey, wet wheat distillers’ grain or
water at different temperatures. Anim. Feed Sci. Technol. 144, 137–148.
Magan N. & Lacey J., 1984. The effect of gas composition and water activity of field and
storage fungi and their interactions. Trans. Br. mycol. Soc. 82, 305-314.
Newman, G., 1971. Investigation of a pressure equalising system in sealed silos. Journal and
proceedings of the Institute of Agicultural Engineering, UK, volume 25 no 4, pp 158-160.
Olstorpe, M., Lyberg, K., Lindberg, J.E., Schnürer, J. & Passoth, V., 2008. Population
diversity of yeasts and lactic acid bacteria in pig feed fermented with whey, wet wheat
distillers’ grains or water at different temperatures. Appl. Environ. Microbiol. 74, 1696–1703.
Olstorpe, M., Axelsson, L., Schnürer, J. & Passoth, V., 2010. Effect of starter culture
inoculation on feed hygiene and microbial population development in fermented pig feed
composed of a cereal grain mix with wet wheat distillers’ grain. J. Appl. Microbiol. 108,
129–138.
Ouwehand, A.C., 1998. Antimicrobial components from lactic acid bacteria. In: Salminen, S.,
Von Wright, A. (Eds.), Lactic Acid Bacteria Microbiology and Functional Aspects. Marcel
Dekker Inc., New York, NY, USA, pp. 139–159.
Pick, E., Noren, O. & Nielsen, V., 1989. Energy Consumption and Input Output Relations in
Field Operations. Food and Agricultural Organization of the United Nations, Rome, Italy.
Salovaara, H., 1998. Lactic acid bacteria in cereal based products. In: Salminen, S., Von
Wright, A. (Eds.), Lactic Acid Bacteria Microbiology and Functional Aspects. Marcel
Dekker Inc., New York, NY, USA, pp. 115–137.
Samson, R.A., Hoekstra, E.S. & Frisvad, J.C., 2004. Introduction to Food- and Airborne
Fungi, 7th edition. Centraalbureau voor Schimmelculture, Institute of the Royal Netherlands
Academy of Arts and Sciences, Utrecht, The Netherlands.
56
Proceedings of the 5th Nordic Feed Science Conference
Forage conservation and feed processing
Agro-food industry wastes conversion into safer feed stock by using lactic acid bacteria
V. Krungleviciute1, E. Bartkiene1, J. Kantautaite1, G. Juodeikiene2, J. Damasius2, V.
Baliukoniene1, B. Bakutis1
1
Lithuanian University of Health Sciences, Tilzes str. 18, 47181 Kaunas, Lithuania
2
Kaunas University of Technology, Radvilenu str. 19, 50254 Kaunas, Lithuania
Corresponding author: [email protected]
Introduction
Wheat and barley are popular products for human consumption. From the manufacture of
them high amount of by-products remain, which are good source for animal feeding. A higher
economical advantage could be achieved with the use of wastes or by-products from agrofood industries, such as the flour milling or groat industries. These products are potential
resources of new safer feed stocks with high nutritional and biological values.
The aim of this study was to evaluate the possibilities of an alternative substrate for
Pediococcus acidilactici multiplication and to adapt multiplied strains for wheat and barley
cereal industry by- product fermentation, as a potential resources to produce safer feed stocks
with a high amount of live lactic acid bacteria (LAB).
Results and Discussion
We found that potato juice could be used as an alternative substrate for P. acidilactici
cultivation. After 72 h of fermentation, a pH of samples 4.1±0.01 was achieved, a total
titratable acidity (TTA) of 8.92±0.06°N (Neiman degrees) and a LAB cell concentration of
9.6±0.07 log10 colony forming units (cfu) per ml.
By spray drying at 150°C or lyophilization (-48°C), the dried powders could be used on
farms as LAB starters for safe feed fermentation because the powders contained viable cell
concentrations of 9.18±0.09 log10 cfu/g and 9.04±0.07 log10 cfu/g, respectively.
Fermentation of cereal by-products (n=7) decreased pH to 4.06±0.01, increased the TTA to
11.9±0.10°N and LAB counts to 9.3±0.3 log10 cfu/g, reduced the occurrence of aflatoxin,
zearalenone, deoxynivalenol and T-2 toxin. Biogenic amine concentrations were changed by
fermentation: phenylethylamine was reduced to 14.9 to 58.2%, putrescine from 46.8 to
97.9%, cadaverine from 69.4 to 100% and spermidine from 14.9 to 58.2% in the fermented
cereal by-products, relative to original concentrations.
Also, P. acidilactici grown on a wheat/barley (10/90; w/w) substrate produced a very small
amount of D-lactic acid and mainly L-lactate. The lactic acid bacteria count of fermented and
dried cereal by-products (n=7) after 90-day storage contained 6.7 log10 cfu/g.
Conclusions
The P. acidilactici multiplied on an alternative substrate such as potato juice can be used for
cereal industry by-product fermentation, as a potential resource to produce safer feed stocks
with high amounts of live lactic acid bacteria with beneficial effects on animal health.
Proceedings of the 5th Nordic Feed Science Conference
57
Ruminant nutrition
Should rumen microbial protein formation be maximized?
G. A. Broderick
Emeritus Professor, University of Wisconsin-Madison and US Dairy Forage Research
Center, Agricultural Research Service-USDA, Madison, Wisconsin 53706, USA
Introduction
Ruminants are efficient users of diets poor in amount and quality of crude protein (CP)
because rumen microbes synthesize the majority of their needs for metabolizable protein
(MP) and essential amino acids (EAA). In addition, microbial utilization of ammonia allows
the feeding of nonprotein N (NPN) compounds, such as urea, plus enables ruminants to utilize
urea-N that is recycled to the rumen (Lapierre and Lobley, 2001). This latter mechanism
captures some N that would otherwise be excreted in the urine. The CP in grass and legume
silages is largely in the form of NPN; microbial protein synthesis converts silage NPN to MP
utilized by the cow. Lactating dairy cows have the highest EAA requirements among
domestic ruminants because of high production of milk protein. Although dairy cattle can
utilize diets with relatively poor quality CP, they typically excrete 2 to 3 times more N in
manure than they secrete in milk. Nitrogen inefficiency necessitates supplementing large
amounts of protein, which increases production costs and contributes to environmental
pollution. Increasing microbial protein formation in the rumen is an important way to improve
N efficiency.
“Optimizing” microbial protein formation in the rumen
The EAA pattern of microbial protein is of better quality than most dietary ingredients
commonly fed to domestic ruminants (Broderick, 1994; Schwab, 1996). Notably, the
proportion of Met and Lys in microbial protein is very similar to that in the proteins in lean
tissue and milk (NRC, 2001). Although there is evidence that His might be limiting in
lactating cows fed diets containing mostly rumen-degraded protein (RDP), animals that
depend largely on microbial protein synthesis (MPS) for their EAA (Korhonen et al., 2000),
most abomasal infusion studies indicate little evidence that a single EAA is limiting under
these conditions (Schwab et al., 2003). However, the total CP in rumen microbial cells
contains from 20% (NRC, 2001) to perhaps 33% (Clark et al., 1992) non-AA N (e.g., N in
nucleic acids and cell-wall components). Thus, conversion of good quality feed proteins into
microbial protein may actually impair N utilization, despite an overall improvement in EAA
pattern, because of a net reduction in EAA supply. There is evidence of linear increases in
MPS in response to increasing dietary RDP (Broderick et al., 2010) while maximal milk
protein secretion occurs at less than maximal MPS (Reynal and Broderick, 2005; Olmos
Colmenero and Broderick, 2006). Hence, MPS should be “optimized” rather than
“maximized”.
Value of NPN in ruminant diets
A number of experiments have shown that ruminants can be moderately productive on diets in
which virtually all CP is supplied as NPN (e.g., Virtanen, 1966). Data from Bryant and
Robinson (1962) indicated that ammonia-N was more important than N from amino acids
(AA) and peptides for growth of many pure cultures of rumen bacteria. However, Oltjen
(1969) summarized results from studies using purified diets showing that replacing CP from
the readily degradable soy protein with that from urea reduced gain, feed efficiency and N
retention in beef steers and heifers. These effects may be partly attributed to rumen escape of
isolated soy protein but one may infer that much of the difference was due to depressed MPS
58
Proceedings of the 5th Nordic Feed Science Conference
Ruminant nutrition
on the urea-based diet. Satter and Slyter (1974) fed diets to mixed rumen organisms in
continuous culture fermenters in which CP content was increased above 4% (DM basis) by
adding only urea. In 3 experiments, ammonia concentrations remained at ~1 mM, and
microbial protein yield increased linearly, until dietary CP reached ~13%. At that point,
protein outflow from the fermenters stopped increasing and ammonia concentration climbed
rapidly; microbial protein yield did not increase above a mean ammonia concentration of ~2
mM (Satter and Slyter, 1974). This value was adjusted to about 3.6 mM (5 mg ammoniaN/dL) for a safety margin. Schaefer et al. (1980) found that ≤1 mM ammonia (1.4 mg N/dL)
gave 95% of maximum growth for 9 of 10 pure cultures of rumen bacteria studied. Results
reported in both studies called into question the value of feeding NPN in many situations.
There has been much debate during the intervening years about whether 5 mg N/dL was in
fact the upper limit for ammonia utilization in vivo. For example, Mehrez et al. (1977) infused
urea into the sheep rumen and found that in situ barley DM digestion increased with
increasing ammonia until reaching ~20 mg ammonia-N/dL. Odle and Schaefer (1987)
conducted similar studies in cattle and observed maximal rates of in situ DM disappearance at
12 and 6 mg ammonia-N/dL for barley and maize. Rumen concentration of diaminopimelic
acid, now a rarely used marker for bacterial protein, increased with urea addition to a high
maize diet until ammonia-N reached 8.5 mg/dL (Kang-Meznarich and Broderick, 1980).
Higher optimal ammonia concentrations may be required under some circumstances because
the physical association of bacteria with particulate matter results in low ammonia levels in
these localized niches (Odle and Schaefer, 1987). Moreover, rumen ammonia varies greatly,
so a mean 5 mg ammonia-N/dL over the day will be the resultant a wide a range of
concentrations (Dixon, 1999).
Amino-N stimulates microbial protein formation
Ammonia is formed partly from deamination of AA derived from rumen protein degradation
and ammonia production parallels formation of peptides and free AA. This confounds the
explanation of whether MPS is responding to ammonia alone, or also to the peptides and AA
that also gave rise to ammonia on natural diets. Work by Lee Baldwin’s group at UC-Davis
shed considerable light in this area with a number of in vitro studies. Maeng et al. (1976)
found that replacing 25% of urea-N with N from a mixture of 18 protein AA maximized the
yield of microbial DM per unit carbohydrate fermented on glucose, starch and cellobiose;
greater proportions of AA-N appeared to reduce microbial yields. Argyle and Baldwin (1989)
showed that adding only 1 mg/L of a blend of protein AA plus 1 mg/L of peptides from
trypsin digested casein (trypticase) more than doubled in vitro cell yield of mixed rumen
organisms. They also found progressively lower response to 10 and 100 times greater addition
of AA and peptides. Furthermore, Argyle and Baldwin (1989) found that mixtures of different
classes of AA did not increase bacterial cell yield when added to a medium containing
ammonia and peptides (trypticase); greater cell yield was only obtained with addition of a
complete mixture of the protein AA. Other data indicated that in vitro MPS increased when
degradation rate of protein sources in the medium increased from 0.01 to 0.14/h, but altered
little with further increase in rate up to 0.54/h (Hristov and Broderick, 1994).
Dramatic responses to RDP supplementation also have been observed in vivo. Chikunya et al.
(1996) found that replacing urea CP with casein CP in sheep fed a grass hay diet (with rumen
OM digestibility of ~32%) increased MPS 10%, but the same replacement in a sugar beet
pulp diet (with rumen OM digestibility of ~51%) increased MPS 82%. Kalscheur et al. (2003)
held rumen-undegraded protein (RUP) supply constant but increased RDP by replacing
Proceedings of the 5th Nordic Feed Science Conference
59
Ruminant nutrition
treated soybean meal with solvent soybean meal; they observed significant increases in yield
of milk, fat and protein at equal DM intake, although N efficiency declined. Recent in vivo
results demonstrated significant linear depression in yield of milk, fat and protein when RDP
from urea replaced RDP from soybean meal; these production effects appeared to be caused
largely by depressed rumen outflow of non-ammonia N (NAN), essential AA and total AA
due to reduced efficiency of microbial protein synthesis (Broderick and Reynal, 2009).
Kozloski et al. (2000) similarly observed a linear decline in MPS as more and more soybean
meal CP was replaced by urea CP. On the other hand, Ahvenjärvi et al. (2002) attributed the
increased flow of NAN to elevated RUP when they supplemented rapeseed meal to dairy
cows fed a grass silage diet. But overall, at least under some conditions, NPN sources cannot
provide all of the RDP and RDP from true protein is required to optimize microbial protein
formation in the in vivo rumen.
Stouthamer (1973) computed theoretical energetic efficiencies of microbial cell growth based
on known metabolic pathways and estimated yields of 28.8 or 28.6 g OM/mol of ATP for
organisms growing on media with N coming from, respectively, ammonia only or AA.
However, Cruz Soto et al. (1994) reported that adding AA or peptides to the medium
increased specific growth rates 2 to 44 fold for several pure cultures of rumen bacteria. Based
on several in vitro studies, Russell et al. (1992) estimated that addition of AA mixtures to
rumen incubations would improve cell yields by 18%, which seems more consistent with in
vivo observations. Wallace’s group at the Rowett Research Institute has explored possible
mechanisms for these increased yields. Adding AA mixtures, and particularly mixed peptides
(trypticase), resulted in substantial reduction in proportions of individual AA in microbial
protein that derived from de novo synthesis (Atasoglu et al., 2004). Deletion of single AA
from a complete mix of protein AA slowed gas production when microbes were grown on a
mixture soluble carbohydrates (Atasoglu et al., 2003) or xylan (Guliye et al., 2005), but did
not alter protein yield in either study. It is not clear what causes the disparity between
experimental observations and theoretical estimates and whether the AA and peptides present
in silages would have similar affects. Russell (2007) speculated that AA and peptides reduced
“energy spilling” during cell growth. The Wallace group (Cruz Soto et al., 1994) concluded
that degree of stimulation by AA and peptides was related to carbohydrate fermentation rate,
with greater responses occurring with more rapidly fermented substrates.
Value of silage CP for microbial protein formation
Charmley and Veira (1990) suppressed proteolysis in ensiled alfalfa using a 2-min steamtreatment, which reduced silage NPN from 65 to 40% of total N without altering NDF or
ADIN. When this alfalfa was fed to sheep, flow of protein N from the rumen increased from
22 to 27 g/d; about 60% of this increase was due to more microbial protein. These workers
also reported that growth efficiency increased from 22 to 37 g microbial NAN/kg OM
apparently digested in the rumen. Thus, breakdown of the forage protein in the silo clearly
depressed microbial protein formation. Formic acid addition to direct-cut, hay-crop silages is
used widely in the U.K. but has been thought unnecessary for forages wilted to greater than
35% DM, such as those commonly harvested in North America. However, Nagel and
Broderick (1992) found that applying formic acid to alfalfa wilted to 40% DM reduced silage
NPN by one-third and increased milk yield 3.4 kg/day and protein yield 110 g/day when fed
to early lactation cows. The results of Charmley and Veira (1990) suggested that the improved
production observed by Nagel and Broderick (1992) was at least partly due to improved
microbial protein formation in the rumen. Alfalfa harvested in 3 trials from the same swards
60
Proceedings of the 5th Nordic Feed Science Conference
Ruminant nutrition
as either 40% DM silage or hay in small rectangular bales averaged 2.5% less CP (20.6 versus
18.1%) when harvested as hay (Broderick, 1995; Vagnoni and Broderick, 1997). However,
despite leaf protein losses, feeding forage as hay rather than silage resulted in greater milk
protein secretion and little or no response to supplementing with RUP. In vitro studies
conducted with samples of forages collected during these 3 in vivo trials indicated that alfalfa
hay and silage had similar RDP and RUP contents but that hay gave rise to greater in vitro
yields of microbial protein (Peltekova and Broderick, 1996). We speculated that similar
effects would happen with grass silages but evidence from the literature does not support this
contention. Protein utilization in cows was not greater on hay compared to silage from the
same sward (Shingfield, 2002). Murphy et al. (2000) also reported similar milk and protein
yields in cows fed either timothy-meadow fescue silage or hay.
Supplementing with fermentable energy
What Murphy et al. (2000) did observe was a trend for increased protein yield, and a highly
significant increase in milk protein concentration, with greater dietary starch content on both
grass silage and hay. Feeding fermentable carbohydrates to match silage RDP supply will
stimulate microbial protein formation and improve N utilization. However, there are
substantial differences among starch sources (Herrera-Saldana et al., 1990), and within grains
due to processing, in the rates of energy release in the rumen. Effects of processing on extent
of rumen digestion of maize starch are much greater than on total tract digestibility (Owens et
al., 1986; Owens et al., 1997). We found that grinding high moisture maize through a 1-cm
screen, reducing mean particle size from 4.3 to 1.7 mm, greatly enhanced rates of in vitro
ammonia uptake by rumen microbes; finer grinds (using screens as small as 0.2-cm) did not
further increase ammonia uptake (Ekinci and Broderick, 1997). Feeding the ground high
moisture maize (1.7 mm mean particle size) to lactating cows increased milk and protein yield
by 2.4 and 0.12 kg/day compared to unground high moisture maize. Much the same thing
happens with dry maize, except at smaller particle sizes. Processing dry shelled maize to
reduce mean particle size from 3.5 to 0.6 mm increased rumen starch digestibility from 54 to
70% (Remond et al., 2004). In a large meta-analysis of literature data, Huhtanen et al. (2008)
found that metabolizable energy intake was the main factor explaining milk yield responses in
cows fed grass silage based diets; silage CP content had a small but significant negative affect
on milk yield in this analysis.
There are “optimal” levels of carbohydrate and forage for supporting rumen protein synthesis
and milk production. The Cornell model (Sniffen et al., 1992) predicts reduced formation of
microbial protein per unit of fermented energy when rumen pH falls below 6.2. This effect
would be particularly problematic in ruminants fed high forage diets where NDF contributes
much of the energy. Rumen pH in high producing dairy cows fed large amounts of
fermentable concentrate may remain below 6.0 for much of the day, often averaging well
below 6.2 overall (Ekinci and Broderick, 1997; Weimer et al., 1999). De Veth and Kolver
(2001) reported that, when pH in rumen continuous culture fermenters was reduced from 6.3
to 5.4 for periods of 0, 4, 8, and 12 hours per day, there was a negative linear relationship
between microbial protein yields and the length of time pH was at 5.4. It is important to
confirm in vivo the pH at which high concentrate feeding begins to reduce utilization of
dietary NPN or otherwise depresses microbial protein formation and to quantify the
magnitude of these effects. Depression of fiber digestion at low rumen pH can aggravate the
problem of depressed milk fat yield by reducing formation of acetate (e.g., Dixon and
Stockdale, 1999), a major milk fat precursor. Adding high moisture maize to an 80% alfalfa
Proceedings of the 5th Nordic Feed Science Conference
61
Ruminant nutrition
silage, 20% concentrate diet to increase concentrate to (% alfalfa silage DM/% concentrate
DM) 65/35, 50/50, and 35/65 gave quadratic responses in DM intake and fat corrected milk
(FCM; Valadares et al., 2000). Intake and FCM yield were maximal at 51% concentrate;
maximum fat yield occurred at 43% concentrate. However, milk and protein yield responses
were not quadratic but linear and both were still going up at 35% forage and 65% concentrate.
Moreover, purine derivative excretion in the urine, an indirect measure of rumen protein
formation, also gave a linear response despite low rumen pH and other signs of over-feeding
of concentrate (Valadares et al., 1999). The lactating cow’s demand for energy is substantial
but short-term reversal trials, such as the Valadares studies (1999, 2000), could mask adverse
effects on rumen and animal health occurring with long-term feeding of excessive
concentrate.
In North America and much of Europe, maize silage is commonly fed as a major portion of
the ration. This high energy “forage” is low in CP and, thus, can be fed to dilute hay-crop
forages containing highly degradable protein (such as legume or grass silages). Dhiman and
Satter (1997) replaced 1/3 or 2/3 of the dietary alfalfa silage with maize silage. Compared to
100% of the forage from alfalfa, milk yield was 6% higher over the whole lactation when 2/3
of the dietary forage was alfalfa silage and 1/3 was maize silage; there also were comparable
improvements in N efficiency. Feeding forage as 1/3 alfalfa silage and 2/3 maize silage
slightly reduced production and substantially increased the need for dietary protein
concentrate. Brito and Broderick (2007) assessed the effects of step-wise replacement of
alfalfa silage with maize silage. The greatest improvement in N efficiency, without loss of
production of milk, fat and protein, occurred at about 50% of the forage from alfalfa silage
and 50% from maize silage, mainly because intake declined when dietary alfalfa was reduced.
As discussed, replacing some dietary starch with very rapidly fermenting sugars may enhance
rumen capture of degraded N. Maize starch was replaced with sucrose (Broderick et al.,
2008), or dried molasses or liquid molasses (Broderick and Radloff, 2004), in three separate
feeding studies; basal diets were formulated from alfalfa and maize silages plus high moisture
maize and soybean meal and averaged 2.6% total sugars in dietary DM. An overall analysis of
the data from the three trials indicated maxima for total sugars (on a DM basis) were 6.8% for
DM intake and 4.8% for protein yield. However, the positive production effects of sugar
feeding in these three trials were primarily driven by increased feed intake.
Summary and conclusions
Dairy cows excrete substantially more N in manure than they secrete in milk, increasing milk
production costs and environmental N pollution. Although the AA pattern of microbial
protein produced in the rumen is of good quality, 20 to 33% of the CP equivalent in microbial
cells is in the form of non-AA N (such as nucleic acids). Optimizing (as opposed to
maximizing) microbial protein formation, by providing the appropriate N compounds in the
rumen, will improve the protein status of the lactating cows and other productive ruminants.
NPN can replace only part of the dietary RDP because RDP from peptides and AA stimulates
microbial protein synthesis (both in amount and efficiency of protein formed per unit of
energy fermentation). The effects of providing RDP in the form of AA and/or peptides are
complex and the mechanism of these improvements in protein yield is not clear. Reducing
dietary NPN by suppressing protein degradation in the silo also improves efficiency of
microbial protein formation. Providing fermentable energy as processed grain and whole-crop
grain silages (e.g., maize silage) is an effective way to improve rumen microbial protein
62
Proceedings of the 5th Nordic Feed Science Conference
Ruminant nutrition
formation from silage NPN. This increases utilization of silage CP and animal performance
while reducing N losses to the environment.
References
Ahvenjarvi, S., A. Vanhatalo & P. Huhtanen. 2002. Supplementing barley or rapeseed meal to
dairy cows fed grass-red clover silage: I. Rumen degradability and microbial flow. J. Anim.
Sci. 80, 2176–2187.
Argyle, J. L. & R. L. Baldwin. 1989. Effects of amino acids and peptides on rumen microbial
yields. J. Dairy Sci., 72, 2017-2027.
Atasoglu, C., A. Y. Guliye & R. J. Wallace. 2003. Use of a deletion approach to assess the
amino acid requirements for optimum fermentation by mixed micro-organisms from the sheep
rumen. Anim. Sci. 76, 147-153.
Atasoglu, C., A. Y. Guliye & R. J. Wallace. 2004. Use of stable isotopes to measure de novo
synthesis and turnover of amino acid-C and -N in mixed micro-organisms from the sheep
rumen in vitro. Brit. J. Nutr. 91, 253–261.
Brito, A. F. & G. A. Broderick. 2007. Effects of different protein supplements on milk
production and nutrient utilization in lactating dairy cows. J. Dairy Sci. 90, 1816-1827.
Broderick, G. A. 1994. Quantifying forage protein quality. In G. C. Fahey, Jr., M. D. Collins,
D. R. Mertens and L. E. Moser (ed.) Forage Quality, Evaluation, and Utilization. American
Soc. Agron., Madison, WI, pp. 200-228.
Broderick, G. A. 1995. Performance of lactating dairy cows fed either alfalfa silage or alfalfa
hay as the sole forage. J. Dairy Sci. 78, 320-329.
Broderick, G. A. & W. J. Radloff. 2004. Effect of molasses supplementation on the
production of lactating dairy cows fed diets based on alfalfa and corn silage. J. Dairy Sci. 87,
2997-3009.
Broderick, G. A. & S. M. Reynal. 2009. Effect of source of rumen-degraded protein on
production and ruminal metabolism in lactating dairy cows. J. Dairy Sci. 92, 2822-2834.
Broderick, G. A., P. Huhtanen, S. Ahvenjarvi, S. M. Reynal & K. J. Shingfield. 2010.
Quantifying ruminal nitrogen metabolism using the omasal sampling technique in cattle--A
meta-analysis. J. Dairy Sci. 93, 3216-3230.
Broderick, G. A., N. D. Luchini, S. M. Reynal, G. A. Varga & V. A. Ishler. 2008. Effect on
production of replacing dietary starch with sucrose in lactating dairy cows. J. Dairy Sci. 91,
4801-4810.
Bryant, M. P. & I. M. Robinson. 1962. Some nutritional characteristics of predominant
culturable ruminal bacteria. J. Bacteriol. 84, 605-614.
Charmley, E. & D. M. Veira. 1990. Inhibition of proteolysis in alfalfa silages using heat at
harvest, Effects on digestion in the rumen, voluntary intake and animal performance. J.
Anim. Sci. 68, 2042-2051.
Chikunya S., C. J. Newbold, L. Rode, X. B. Chen & R. J. Wallace. 1996. Influence of dietary
rumen-degradable protein on bacterial growth in the rumen of sheep receiving different
energy sources. Anim. Feed. Sci. Tech. 63, 333-340.
Proceedings of the 5th Nordic Feed Science Conference
63
Ruminant nutrition
Clark, J. H., T. H. Klusmeyer & M. R. Cameron. 1992. Microbial protein synthesis and flows
of nitrogen fractions to the duodenum of dairy cows. J. Dairy Sci. 75, 2304-2323.
Cruz Soto, R., S. A. Muhammed, C. J. Newbold, C. S. Stewart & R. J. Wallace. 1994.
Influence of peptides, amino acids and urea on microbial activity in the rumen of sheep
receiving grass hay and on the growth of rumen bacteria in vitro. Anim. Feed Sci. Technol.
49, 151-161.
De Veth, M. J. & E. S. Kolver. 2001. Diurnal variation in pH reduces digestion and synthesis
of microbial protein when pasture is fermented in continuous culture. J. Dairy Sci. 84, 20662072.
Dhiman, T. R. & L. D. Satter. 1997. Yield response of dairy cows fed different proportions of
alfalfa silage and corn silage. J. Dairy Sci. 80, 2069-2082.
Dixon, R. M. 1999. Effects of addition of urea to a low nitrogen diet on the rumen digestion
of a range of roughages. Australian J. Agric. Res. 50, 1091-1097.
Dixon, R. M. & C. R. Stockdale. 1999. Associative effects between forages and grains,
Consequences for feed utilization. Australian J. Agric. Res. 50, 757-773.
Ekinci, C. & G. A. Broderick. 1997. Effect of processing high moisture corn on ruminal
fermentation and milk production. J. Dairy Sci. 80, 3298-3307.
Guliye, A. Y., C. Atasoglu & R. J. Wallace. 2005. Assessment of amino acid requirements for
optimum fermentation of xylan by mixed micro-organisms from the sheep rumen. Anim. Sci.
80, 353-360.
Herrera-Saldana, R., R. Gomez-Alarcon, M. Torabi & J. T. Huber. 1990. Influence of
synchronizing protein and starch degradation in the rumen on nutrient utilization and
microbial protein synthesis. J. Dairy Sci. 73, 142-148.
Hristov, A. & G. A. Broderick. 1994. In vitro determination of ruminal protein degradability
using [15N]-ammonia to correct for microbial nitrogen uptake. J. Anim. Sci. 72, 1344-1353.
Huhtanen, P., M. Rinne & J. Nousiainen. 2008. Effects of silage soluble nitrogen components
on metabolizable protein concentration, A meta-analysis of dairy cow production
experiments. J. Dairy Sci. 91, 1150-1158.
Kalscheur, K. F., R. L. Baldwin VI, B. P. Glenn & R. A. Kohn. 2003. Milk production of
dairy cows fed differing concentrations of rumen-degraded protein. J. Dairy Sci. 89, 249-259.
Kang-Meznarich, J. H. & G. A. Broderick. 1980. Effects of incremental urea supplementation
on ruminal ammonia concentration and bacterial protein formation. J. Anim. Sci. 51, 422-431.
Korhonen, M., A. Vanhatalo, T. Varvikko & P. Huhtanen. 2000. Responses to graded
postruminal doses of histidine in dairy cows fed grass silage diets. J. Dairy Sci. 83, 25962608.
Kozloski, G. V., H.M.N. Ribeiro-Filho & J.B.T. Roch. 2000. Effect of the substitution of urea
for soybean meal on digestion in steers. Canadian J. Anim. Sci. 80, 713-719.
Lapierre, H. & G. E. Lobley. 2001. Nitrogen recycling in the ruminant. A review. J. Dairy
Sci. 84(E. Suppl.), E223-E236.
64
Proceedings of the 5th Nordic Feed Science Conference
Ruminant nutrition
Maeng, W. J., C. J. Van Nevel, R. L. Baldwin & J. G. Morris. 1976. Rumen microbial
growth rates and yields. Effect of amino acids and protein. J. Dairy Sci. 59, 68-79.
Mehrez, A. Z., E. R. Ørskov & I. McDonald. 1977. Rates of rumen fermentation in relation to
ammonia concentration. Brit. J. Nutr. 38, 437-443.
Murphy, M., M. Åkerlind & K. Holtenius. 2000. Rumen fermentation in lactating cows
selected for milk fat content fed two forage to concentrate ratios with hay or silage. J. Dairy
Sci. 83, 756-764.
Nagel, S.A. & G.A. Broderick. 1992. Effect of formic acid or formaldehyde treatment of
alfalfa silage on nutrient utilization by dairy cows. J. Dairy Sci. 75, 140-154.
National Research Council. 2001. Nutrient Requirements of Dairy Cattle. 7th rev. ed. Natl.
Acad. Sci., Washington, DC.
Odle, J. & D. M. Schaefer. 1987. Influence of rumen ammonia concentration on the rumen
degradation rates of barley and maize. Brit. J. Nutr. 57, 127-138.
Olmos Colmenero, J. J. & G. A. Broderick. 2006. Effect of dietary crude protein
concentration on milk production and nitrogen utilization in lactating dairy cows. J. Dairy Sci.
89, 1704-1712.
Oltjen, R. R. 1969. Effects of feeding ruminants non-protein nitrogen as the only nitrogen
source. J. Anim. Sci. 28, 673-682.
Owens, F. N., D. S. Secrist, W. J. Hill & D. R. Gill. 1997. The effect of grain source and grain
processing on performance of feedlot cattle. A review. J. Anim. Sci. 75, 868-879.
Owens, F. N., R. A. Zinn & Y. K. Kim. 1986. Limits to starch digestion in the ruminant small
intestine. J. Anim. Sci. 63, 1634-1648.
Peltekova, V. D. & G. A. Broderick. 1996. In vitro ruminal degradation and synthesis of
protein on fractions extracted from alfalfa hay and silage. J. Dairy Sci. 79, 612-619.
Remond, D., J. I. Cabrera-Estrada, M. Champion, B. Chauveau, R. Coudure & C. Poncet.
2004. Effect of corn particle size on site and extent of starch digestion in lactating dairy cows.
J. Dairy Sci. 87, 1389–1399.
Reynal, S. M. & G. A. Broderick. 2005. Effect of dietary level of rumen-degraded protein on
production and nitrogen metabolism in lactating dairy cows. J. Dairy Sci. 88, 4045-4064.
Russell, J. B. 2007. The energy spilling reactions of bacteria and other organisms. J. Mol.
Microbiol. Biotechnol. 13, 1-11.
Russell, J. B., J. D. O'Connor, D. G. Fox, P. J. Van Soest & C. J. Sniffen. 1992. A net
carbohydrate and protein system for evaluating cattle diets. I. Ruminal fermentation. J. Anim.
Sci. 70, 3551-3561.
Satter, L. D. & L. L. Slyter. 1974. Effect of ammonia concentration on rumen microbial
protein production in vitro. Brit. J. Nutr. 32, 199-208.
Schaefer, D. M., C. L. Davis & M. P. Bryant. 1980. Ammonia saturation constants for
predominant species of rumen bacteria. J. Dairy Sci. 63, 1248-1263.
Schwab, C. G. 1996. Rumen-protected amino acids for dairy cattle. Progress towards
determining lysine and methionine requirements. Anim. Feed Sci. Tech. 59, 87-101.
Proceedings of the 5th Nordic Feed Science Conference
65
Ruminant nutrition
Schwab, C. G., R. S. Ordway & N. L. Whitehouse. 2003. Amino acid balancing in the context
of MP and RUP requirements. Pages 10-25 in Proc. Four-State Applied Nutrition and
Management Conference.
Shingfield K. J., S. Jaakkola & P. Huhtanen. 2002. Effect of forage conservation method,
concentrate level and propylene glycol on intake, feeding behavior and milk production of
dairy cows. Anim. Sci. 74, 383–397.
Sniffen, C. J., J. D. O'Connor, P. J. Van Soest, D. G. Fox & J. B. Russell. 1992. A net
carbohydrate and protein system for evaluating cattle diets. II. Carbohydrate and protein
availability. J. Anim. Sci. 70, 3562-3577.
Stouthamer, A. H. 1973. A theoretical study on the amount of ATP required for synthesis of
microbial cell material. Ant. van Leeuwenhoek 39, 545-565.
Vagnoni, D. B. & G. A. Broderick, G. A. 1997. Effects of supplementation of energy or
ruminally undegraded protein to lactating cows fed alfalfa hay or silage. J. Dairy Sci. 80,
1703-1712.
Valadares, R.F.D., G. A. Broderick, S. C. Valadares Filho & M. K. Clayton. 1999. Effect of
replacing alfalfa silage with high moisture corn on ruminal protein synthesis estimated from
excretion of total purine derivatives. J. Dairy Sci. 82, 2686-2696.
Valadares Filho, S. C., G. A. Broderick, R.F.D. Valadares & M. K. Clayton. 2000. Effect of
replacing alfalfa silage with high moisture corn on nutrient utilization and milk production. J.
Dairy Sci. 83, 106-114.
Virtanen, A. I. 1966. Milk production of cows on protein-free feed. Science 153, 1603-1614.
Weimer, P. J., G. C. Waghorn, C. L. Odt & D. R. Mertens. 1999. Effect of diet on populations
of three species of ruminal cellulolytic bacteria in lactating dairy cows. J. Dairy Sci. 82, 122134.
66
Proceedings of the 5th Nordic Feed Science Conference
Ruminant nutrition
Biological limits to N utilization in dairy cows
S. Ahvenjärvi1, T. Stefanski1, A. Vanhatalo2 & P. Huhtanen3
1
MTT Agrifood Research Finland, Animal Production, 31600 Jokioinen, Finland,2University
of Helsinki, Department of Agricultural Sciences, P.O. Box 28, 00014 University of Helsinki,
Finland,3 Swedish University of Agricultural Sciences (SLU), Department of Agricultural
Research for Northern Sweden,S-901 83 Umeå, Sweden.
Correspondence: [email protected]
Introduction
Dairy cow diets are often supplemented with protein rich ingredients to satisfy alleged protein
requirements. Increases in dietary crude protein (CP) concentration often increase milk yield
but consistently decrease the efficiency of N use for milk protein synthesis (Broderick, 2003;
Ipharraguerre and Clark, 2005). The conflict of interest between economically profitable and
environmentally friendly decision-making was clearly highlighted in a recent meta-analysis
by Huhtanen et al. (2011). Daily milk yield responses to supplementary protein were between
2.1 to 3.7 kg kg-1 increase in CP intake, whereas the responses in milk protein yield were only
between 98 to 136 g kg-1 increase in CP intake. These results suggest that while milk yield
responses may be economically profitable they are biologically very inefficient because in
average case 86 to 90% of high quality feed protein will be lost as N in urine and faeces. In
Finland, over the next few years dairy business will be facing increasing pressures to improve
the efficiency of N use because Finland has committed to reduce ammonia emissions by 20%
relative to year 2005 by the year 2020. Based on substantial experimental evidence by far the
most efficient means to improve N use efficiency is to reduce dietary CP concentration
(Broderick, 2003; Huhtanen and Hristov, 2009). The objective of the current paper is to
present some previous findings on ruminal N metabolism underlying N utilization for milk
protein synthesis.
Utilization of protein vs. non-protein N
In our previous study (Stefanski et al., unpublished), ruminal metabolism of ammonia-N and
feed protein were studied using 15N labelled N sources. Labelled ammonium-N, soluble and
insoluble rapeseed meal protein were introduced as a pulse dose into the rumen of lactating
dairy cows. Cumulative secretion of 15N in milk over the course of 108 h post dose indicated
that 19, 20, and 22% of labelled N administered as ammonium-N, soluble and insoluble
rapeseed meal protein was utilized for milk protein synthesis (MNE15N), respectively. These
differences between non-protein-N, soluble protein and insoluble protein were smaller than
expected and could be attributed to moderate CP concentration in the diet (155 g/kg DM) and
low ruminal ammonium-N concentration (6.0 mg/100 mL). Low dietary CP and ruminal
ammonia-N concentrations favour efficient utilization of rumen degradable N for microbial
protein synthesis and result in small net absorption and outflow of ammonia-N from the
rumen. In addition, the efficiency of N recycling as urea into the rumen increases with
decreases in dietary CP concentration (Marini and Amburgh, 2003). On a low CP diet (91
g/kg DM) approximately 43% of the N recycled into the gastrointestinal tract was used for
microbial N synthesis, whereas only 6% was utilized on a high CP diet (213 g/kg DM; Marini
and Amburgh, 2003).
In our study, the efficiency of N use (milk N/N intake) determined for the whole diet was
32% (MNEdiet). The difference between MNEdiet and MNE15N could be attributed to
deposition of labelled N in tissue protein that turns over slower than 108 h period of milk
Proceedings of the 5th Nordic Feed Science Conference
67
Ruminant nutrition
protein synthesis accounted for. Following this hypothesis MNE15N should indicate the extent
of milk protein that was synthesized either directly or via rapid turnover body tissue pools,
whereas the difference between MNEdiet and MNE15N should indicate the proportion that was
derived from slow turnover body tissues. Based on such comparison, MNE15N accounted for
59 to 69% of MNEdiet suggesting that between 60 to 70% of milk protein was synthesized
from amino acids absorbed from the gastrointestinal tract, whereas 30 to 40 % of milk protein
was derived from body tissues.
The relationship between cumulative secretion of 15N labelled ammonia-N in milk as a
proportion of a single dose administered into the rumen of lactating dairy cows (Figure 1),
compiled from three separate studies, points out a substantial effect of dietary CP
concentration on the efficiency of N utilization. In ruminants, the efficiency of urea recycling
and the proportion of recycled N used for microbial protein synthesis are heavily dependent
on dietary CP concentration (Marini and Van Amburgh, 2003). Studies on the renal function
of sheep have indicated that animal physiology has an important role in conservation of N
(Leng et al., 1985). Sheep fed a low N diet retained urea by the mechanism of decreases in
renal glomerular filtration rate and increases in tubular transport capacity for urea (Leng et al.,
1985).
Cumulative secretion in milk, %
20
19
18
17
16
15
y = -1.24x + 37.8
R² = 0.90
14
13
12
15
16
17
18
19
20
Diet CP%
Figure 1 Cumulative secretion of 15N labelled ammonia-N in milk as a proportion of a single dose introduced
into the rumen. Data from Hristov and Ropp, 2003; Hristov et al., 2004; Stefanski et al. unpublished.
Critical rumen ammonia-N concentration
Our previous study (Ahvenjärvi and Huhtanen, unpublished) was designed to identify the
critical dietary CP and rumen ammonia levels below which ruminal fiber digestibility and
microbial N synthesis are decreased relative to optimal conditions. Dairy cows were offered a
68
Proceedings of the 5th Nordic Feed Science Conference
Ruminant nutrition
total mixed ration that was formulated slightly deficient in terms of rumen degradable protein
(CP 127 g/kg DM, PBV -3 g/kg DM). The basal diet was supplemented with increasing levels
or urea infusion (17, 33, 49 and 66 g/d of N) into the rumen to increase ammonia-N supply to
rumen microbes. On the basal diet, average rumen ammonia-N and milk urea concentrations
remained rather low 2.7 mmol/l and 9.5 mg/100 ml, respectively. In spite of low rumen
ammonia-N concentrations cows didn’t exhibit responses to intraruminal urea infusions in
terms of feed intake, NDF digestibility or energy corrected milk yield. However, increases in
milk protein yield (P = 0.08) and faecal N excretion (P = 0.05) indicated that microbial N
synthesis slightly increased as a response to urea infusion. Interestingly, rumen ammonia-N
concentration didn’t exhibit any pattern of increase until the highest level (66 g/d) was
achieved. This inflection point in rumen ammonia-N concentration was achieved at a dietary
CP concentration of 140 g/kg DM. In contrast to rumen ammonia, milk urea concentrations
increased already on the lowest levels of urea infusion (Figure 1). Lack of response in rumen
ammonia-N concentration may be explained by active ammonia transportation into microbial
cells with subsequent ammonia-N gradient between intracellular and extracellular space
(Russell and Strobel, 1987). Therefore, rumen ammonia-N concentration may not be a
satisfactory indicator of N deficiency in the rumen.
The relationship between urea infusion levels and increases in N excretion indicated that 0.67
(R2 = 0.99) of urea-N was excreted in urine, 0.14 (R2 = 0.69) in feces and 0.22 (R2 = 0.78)
was secreted in milk. These estimates represent marginal utilization of supplementary
ammonia-N provided as continuous infusion into the rumen. Compared with marginal
utilization of soybean meal protein for milk protein synthesis (98 g kg-1), reported by
Huhtanen et al. (2011), the marginal utilization of ammonia-N in the current study was
approximately two times higher. This indirect comparison clearly points out that N utilization
is more dependent on dietary CP concentration than the quality of protein. The proportion of
supplementary N utilized for microbial N synthesis is higher than the responses in milk N
output owing to inevitable losses in the intermediary metabolism of absorbed amino acids.
Assuming 0.80 of rumen microbial N entering the intestines is true protein, the digestibility of
this true protein is 0.80, and the efficiency of use of metabolizable protein to milk protein
synthesis is 0.67 (NRC, 2001) the maximum efficiency of microbial protein utilization of
milk protein synthesis would be 0.43. Assuming as in the current study, 0.22 of
supplementary urea-N is transferred to milk N then, calculating back to the proportion used by
the rumen microbes (0.22/0.43), this suggests that 0.51 of urea-N was captured by rumen
microbes. This figure represents both immediate capture of urea-N to microbial N and also
recycled ammonia-N returned to the forestomach via hepatic urea synthesis. This estimate is
in close agreement with the mean values for urea-N kinetics estimated by Lapierre and Lobley
(2001). These authors estimated that typically 0.33 of hepatic urea flux is eliminated in urine,
whereas 0.67 enters the digestive tract. Of the latter, 0.50 is reabsorbed as anabolic-N
(microbial N), whereas 0.40 is reabsorbed as ammonia-N and 0.10 is lost in feces.
Proceedings of the 5th Nordic Feed Science Conference
69
Ruminant nutrition
3.8
18.0
3.6
17.0
Rumen ammonia, mmol/l
15.0
3.2
14.0
3.0
13.0
2.8
12.0
2.6
11.0
2.4
Milk urea, mg/100 ml
16.0
3.4
10.0
2.2
9.0
2.0
8.0
0.0
10.0
20.0
30.0
40.0
50.0
60.0
70.0
Urea-N, g/d
RAN
MU
Figure 2 The effect of intraruminal urea-N infusion on rumen ammonia-N (RAN) and milk urea (MU)
concentration.
Conclusions
Observations from our previous studies demonstrate that a lactating cow is capable of
efficient utilization of both protein and non-protein N provided that dietary CP concentration
doesn’t exceed microbial requirements. However, more experimental evidence is needed to
account for the contribution of urea recycling to ruminal N supply in the future models.
References
Broderick, G.A., 2003. Effects of varying dietary protein and energy levels on the production
of lactating dairy cows. J. Dairy Sci., 86, 1370-1381.
Hristov, A.N., Etter, R.P., Ropp, J.K. & Grandeen, K.L., 2004. Effect of dietary crude protein
level and degradability on ruminal fermentation and nitrogen utilization in lactating dairy
cows. J. Anim. Sci., 82, 3219-3229.
Hristov, A.N. & Ropp, J.K., 2003. Effect of dietary carbohydrate composition and availability
on utilization of ruminal ammonia nitrogen for milk protein synthesis in dairy cows. J. Dairy
Sci., 86, 2416-2427.
Huhtanen, P.H., Hetta, M. & Swensson, C., 2011. Evaluation of canola meal as a protein
supplement for dairy cows: a review and a meta-analysis. Can. J. Anim. Sci., 91, 529-543.
70
Proceedings of the 5th Nordic Feed Science Conference
Ruminant nutrition
Huhtanen, P.J. & Hristov, A.N., 2009. A meta-analysis of the effects of dietary protein
concentration and degradability on milk protein yield and milk N efficiency in dairy cows. J.
Dairy Sci. 92, 3222-3232.
Ipharraguerre, I.R. & Clark, J.H., 2005. Varying protein and starch in the diet of dairy cows.
II. Effects on performance and nitrogen utilization for milk production. J. Dairy Sci., 88,
2556-2570.
Lapierre, H. & Lobley G.E., 2001. Nitrogen recycling in the ruminant: a review. J. Dairy Sci.,
84, E223-E236.
Leng, L., Szanyiova, M. & Boda, K., 1985. The renal response of sheep to a low dietary
nitrogen intake. Physiologia Bohemoslovaca, 34, 147-154.
Marini, J.C. & Van Amburgh, M.E., 2003. Nitrogen metabolism and recycling in Holstein
heifers. J. Anim. Sci., 81, 545-552.
NRC., 2001. Nutrient Requirements of Dairy Cattle. 7th rev. ed. Natl. Acad. Sci.,
Washington, DC.
Russell, J.B. & Strobel, H.J., 1987. Concentration of ammonia across cell membranes of
mixed rumen bacteria. J. Dairy Sci., 70, 970-976.
Proceedings of the 5th Nordic Feed Science Conference
71
Ruminant nutrition
Evaluation of protein supplementation for growing cattle: A meta-analysis
A. Huuskonen1, P. Huhtanen2 & E. Joki-Tokola1
1
MTT Agrifood Research Finland, Animal Production Research, Tutkimusasemantie 15, FI92400 Ruukki, Finland. 2 Swedish University of Agricultural Sciences (SLU), Department of
Agricultural Research for Northern Sweden, S-90183 Umeå, Sweden
Correspondence: [email protected]
Introduction
Ruminants have two types of N requirements, the N requirements of ruminal fermentation and
the amino acid (AA) requirements of the host animal. A shortage of rumen degradable feed
protein (RDP) has been shown to reduce microbial digestion of carbohydrates, reduce
synthesis of microbial protein, decrease feed intake and decrease weight gain of growing
cattle (Schwab et al. 2005). A shortage of absorbed AA by cattle, either because of decreased
synthesis of microbial protein or less than required intakes of rumen-undegraded protein
(RUP), decrease weight gain of growing cattle (Lammers & Heinrichs 2000). Even though
individual studies have examined the influence of protein supplementation on the intake and
performance of growing cattle, limitations in the number of observations and basal diet
composition do not allow for definitive conclusions. Therefore, the objective of this metaanalysis was to develop empirical equations predicting performance responses of growing
cattle to protein intake.
Materials and Methods
A dataset was collected from published feeding experiments with growing cattle (bulls, steers,
heifers) fed ad libitum total mixed ration (TMR) or ad libitum grass silage, or whole-crop
silages (barley, oats or wheat), hay or straw partly or completely substituted for grass silage
fed with fixed amounts of concentrates. The concentrate feeds consisted of cereal grains,
fibrous by-products and various protein supplements [mainly rapeseed meal (RSM), soybean
meal (SBM) and fish meal (FM)]. Approximately half of the studies were conducted with
pure dairy breeds and the remainder with beef breeds or dairy × beef crossbred animals.
Approximately half of the experiments were conducted in the UK and Ireland and the
remainder in the Nordic countries. Variation in the design of experiments, animal
performance, experimental diets, feeding routines, etc. between the experiments was
substantial covering most practical on-farm feeding alternatives. The minimum prerequisite
for an experiment to be included in the dataset was that forage and total dry matter (DM)
intakes and initial and final body weights (BW) were available and adequate forage
characterization [plant species, DM and crude protein (CP) concentration, in vivo or in vitro
digestibility of organic matter and fermentation quality] and adequate concentrate
characterization (proportion of ingredients, DM, CP concentration) were available. Overall,
the dataset comprised 199 diets in 80 studies. More details and a complete reference list are
provided in a paper by Huuskonen et al. (2014).
The dataset was analysed to evaluate the effects of different dietary variables on body weight
gain (BWG) and carcass traits with a MIXED model regression analysis of SAS: Y = B0 +
B1X1ij + b0 + b1X1ij + eij, where B0 and B1X1ij are the fixed effects (intercept and effects of
independent variables) and b0, b1, and eij are the random experiment effects (intercept and
slope), where i = 1 . . . n studies and j = 1 . . . ni values. Unstructured variance-covariance
matrix for the intercepts and slopes (TYPE = UN option) was used in the random statement.
The significance of the differences in the slope of dependent variable was tested with t-test.
72
Proceedings of the 5th Nordic Feed Science Conference
Ruminant nutrition
Root mean squared errors (RMSE) presented in the tables are adjusted for random study
effects as described by St-Pierre (2001). The effects of different animal and dietary variables
on BWG responses to increased dietary CP concentration were evaluated by plotting the
random slope effect (B1 + b1) in each study against observed animal and diet variables on the
control diet in each study (lowest dietary CP concentration). The responses to increased
dietary metabolisable protein (MP) concentration were also analysed but because the trends
were similar as in CP the results were not presented.
Results and Discussion
Chemical composition of feeds and total diets as well as intake and performance data were
reported by Huuskonen et al. (2014). The forage and concentrate components and the total
diets displayed wide ranges in chemical composition and feeding values. On average, the
silages were well fermented, but the dataset also included both extensively and poorly
fermented silages. Total DM intake (DMI) ranged from 3.3 to 10.4 kg/d reflecting, for
example, differences in BW of the animals and the intake potential of the diets. The large
variation in BWG (from 521 to 1809 g/d) reflects differences in genetic potential of the
animals and nutritive value and intake potential of the diets.
Increasing dietary CP concentration increased BWG (P<0.01; Table 1). However, the
responses were quantitatively small (1.4 g BWG per 1 g/kg DM increase in dietary CP
concentration, on average). The BWG responses were not different for bulls vs. steers and
heifers (1.4 and 1.3 g per 1 g/kg DM increase in dietary CP concentration) and for dairy and
beef breeds (1.2 vs. 1.7 g per 1 g/kg DM, respectively). The response showed diminishing
responses with increased CP concentration (negative quadratic effect P<0.01). In terms of
Akaike’s information criteria (AIC) the MP model performed slightly better than the CP
model (Table 1). When RUP and RDP were used in a bivariate model, only RUP had a
positive effect on BWG and the slope was markedly greater than for CP. The BWG response
to increased dietary CP concentration increased with decreased effective protein degradability
(EPD). In terms of the smaller AIC the intake models performed better than the models based
on protein concentrations. The model based on ME intake, CP concentration and EPD was the
best on the basis of the smallest AIC.
The effect of increased diet CP concentration on BWG declined (P<0.01) with increasing
mean BW of the animals and with improved BWG of the control animals (the lowest CP diet
in each study) (Fig. 1). The BWG responses to increased protein supplementation were
negatively related to DMI (P<0.01) and ME intake (P<0.01) (data not shown). This indicates
that protein can be more limiting when diet characteristics are limiting intake and synthesis of
microbial protein. Surprisingly, BWG responses to protein supplementation were not related
to the CP concentration in the diet of animals fed the control diet (Fig. 1). The responses
increased (P<0.05) with increased ammonia N concentration in silage N, and declined
marginally (P>0.10) with increasing proportion of concentrate in the diet and increasing
dietary concentrations of ME and MP. The BWG responses to protein supplementation were
not related to the digestibility or total acids concentration of the forages (data not shown).
Proceedings of the 5th Nordic Feed Science Conference
73
Ruminant nutrition
Table 1 Relationships between diet/intake parameters and body weight gain
Intercept
SEa
Pvalue
X1
SE
Pvalue
Concentrationd
CP
CP
CP2
921
295
60
224
<0.01
0.19
1.3
9.7
0.34
2.87
<0.01
<0.01
CP
MP
MP
1310
663
-891
225
114
1194
<0.01
<0.01
0.46
1.0
5.1
39
0.38
1.16
25.8
0.01
<0.01
0.14
X-variables
EPD
MP2
X2
SE
-0.028 0.0093
Pvalue
X3
SE
PAdj.
value RMSEb
<0.01
-416
233
0.08
-0.18
0.140
0.20
AICc
32.3
33.3
2423
2423
32.8
37.7
36.6
2407
2415
2416
RUP RDP
1007
56
<0.01 3.2 1.17 <0.01
0.18
0.538
0.74
35.3
2418
Intakee
DMI
510
92
<0.01
84 12.2 <0.01
35.8
2362
MEI
527
88
<0.01 7.2 1.00 <0.01
37.0
2362
CPI
834
53
<0.01 284 46.3 <0.01
28.3
2378
MPI
616
74
<0.01 786 107 <0.01
32.3
2345
MEI CPI
508
88
<0.01 6.0 1.09 <0.01
106
32.4
<0.01
34.8
2343
MEI CP
359
94
<0.01 7.4 1.00 <0.01
1.01
0.229 <0.01
33.7
2345
MEI CP
EPD
826
206 <0.01 7.2 0.98 <0.01
0.65
0.268
0.02 -496 199 0.02
33.7
2327
a
SE = standard error; b Adj. RMSE = Residual mean squared error adjusted for random study effects; c Akaike’s information criteria; d CP = crude protein (g/kg
DM), EPD = effective protein degradability (kg/kg), MP = metabolisable protein (g/kg DM), RUP = rumen undegraded protein (g/kg DM), RDP = rumen degraded
protein (g/kg DM); e DMI = dry matter intake (kg/d), MEI = metabolisable energy intake (MJ/d), CPI = CP intake (kg/d), MPI = metabolisable protein intake
(kg/d).
74
Proceedings of the 5th Nordic Feed Science Conference
Ruminant nutrition
5
5
4
y = -0.0043x + 2.90
R2 = 0.261
BW gain response (g/g)
BW gain response (g/g)
4
3
2
1
0
100
200
300
400
500
600
y = -0.0035x + 5.15
R2 = 0.582
3
2
1
0
400
800
1200
1600
2000
-1
-1
-2
-2
BW gain on control diet (g/d)
Mean BW (kg)
5
5
BW gain response (g/g)
BW gain response (g/g)
4
3
2
1
0
75
100
125
150
175
y = 0.0095x + 0.6883
R2 = 0.0715
4
y = -0.0017x + 1.57
R2 = 0.001
3
2
1
0
0
200
20
40
60
80
100
120
140
160
180
200
-1
-1
-2
-2
Diet CP on control (g/kg DM)
Silage ammonia N (g/kg total N)
Figure 1 The effects of mean BW, BW gain of the control group (the lowest CP in each study), dietary CP concentration of the control group (the lowest CP in
each trial) and silage ammonia N concentration on BW gain response (g/d per 1 g CP/kg DM) to increased CP concentration in the diet.
Proceedings of the 5th Nordic Feed Science Conference
75
Ruminant nutrition
Increasing dietary CP concentration improved (P<0.01) feed efficiency expressed as BWG
per kg DMI or MJ ME intake, whereas BWG per kg CP intake decreased markedly with
increased dietary CP concentration (Table 2). Assuming CP concentration of 170 g/kg BW
the marginal efficiency of the utilization of incremental CP intake was only 0.05 (SE =
0.008). Increased dietary CP concentration had no significant (P>0.10) effects on days in
study, carcass weight, dressing proportion or carcass conformation score, but it increased
(P<0.01) carcass fat score.
The differences between protein variables in predicting BW differences were generally small.
Negative influence of EPD and greater positive effect of RUP compared with RDP suggest
that the small positive responses to supplementary protein were associated with increased
supply of RUP. There was no relationship between dietary CP concentration on the control
diet (the lowest CP in the study) and BWG response to supplementary protein suggesting that
the requirements of RDP were met by all diets or that recycling compensated for the limited
supply from the control diet. The Finnish recommendation for growing cattle above 200 kg
BW is that the protein balance in the rumen (PBV) is above -10 g/kg DM (MTT 2014).
Deleting studies in which PBV of the control diet was below -20 g/kg DM had only a
minimal influence on BWG response to increased CP (1.1 vs. 1.3 g per g CP/kg DM)
suggesting that recommended PBV can even be reduced without adverse effects on BWG.
The amounts of N recycled into the gastrointestinal tract was 27 g/kg DMI in cattle fed a low
CP (80 g/kg DM) diet and approximately 40 g/kg DMI in cattle fed higher CP diets (Marini
& Van Amburgh 2003). These values indicate that in growing cattle rumen PBV can be
negative with minimal, if any, adverse effects on the BWG. Advantages of using MP in
estimating protein supply and requirements are questionable at least for growing cattle above
200 kg BW; microbial protein and RUP from high quality forages and energy supplements
(grain) can meet the requirements.
Table 2
The effects of protein supplementation on feed efficiency and carcass traits
Feed efficiency
BWGd / DMIe (g/kg)
BWG / MEIf (g/MJ)
BWG / CPIg (g/g)
Carcass traits
Days in trial
Carcass weight (kg)
Dressing proportion (g/kg)
Conformation scoreh
Intercept
SEa
P-value
X1
SE
P-value
Adj.
RMSEb
AICc
138
11.6
2.03
8.2
0.70
0.084
<0.01
<0.01
<0.01
0.18
0.019
-0.0061
0.046
0.0041
0.00040
<0.01
<0.01
<0.01
3.9
0.33
0.026
1706
730
-247
237
254
523
6.7
2.7
15.5
19.2
5.8
0.67
0.10
<0.01
<0.01
<0.01
<0.01
<0.01
-0.09
0.15
0.05
0.0009
0.0033
0.058
0.113
0.037
0.00332
0.00064
0.13
0.20
0.15
0.78
<0.01
5.7
3.0
2.7
1875
1442
1225
265
182
0.23
0.18
Fat scorei
a
b
SE = standard error; Adj. RMSE = Residual mean squared error adjusted for random study effects; c Akaike’s
information criteria; d BWG = Body weight gain (kg/d); e DMI = dry matter intake (kg/d); f MEI =
metabolisable energy intake (MJ/d); g CP = crude protein (g/kg DM); h Conformation score: 1 = poorest; 15 =
excellent; i Fat score: 0 = leanest; 10 = fattest.
Assuming a CP concentration of 170 g/kg BW, marginal efficiency of CP utilisation was
0.05. Efficiency of N utilisation (N retention/N intake) decreased by 1.02 (SE=0.05; P<0.01)
g/kg per 1 g/kg DM increase in dietary CP concentration. These numbers indicate a poor
76
Proceedings of the 5th Nordic Feed Science Conference
Ruminant nutrition
utilisation of supplementary protein in growing cattle and consequently, increased N
emissions to the environment. True CP digestibility of RSM and SBM estimated by the Lucas
test was on average 0.95 in dairy cows (Huhtanen et al. 2011). These numbers suggest that
about 90% of incremental N intake would be excreted in urine in growing cattle fed
supplementary protein. Because urinary N is more susceptible to both leaching and
evaporation the adverse effects to the environment increase proportionally more than N
output in manure. Moreover, the concentration of P in protein supplements is usually high,
especially in RSM (12–14 g/kg DM). Consequently, supplementary protein feeding also
increases P emission, since the concentration P in the basal diet based on forages and grain
typically contains 3.0–3.5 g P/kg DM in excess of the requirements of growing cattle (MTT
2014).
Conclusions
Due to quantitatively small production responses, increased feed costs and emissions of N
and P, there is generally no benefit in using protein supplementation to growing cattle fed
grass silage-based diets, provided that the supply of RDP is not limiting digestion in the
rumen. The results suggest no benefits of feed protein systems based on metabolisable protein
for cattle above 200 kg BW.
References
Huhtanen, P., Hetta, M. & Swensson, C., 2011. Evaluation of canola meal as a protein
supplement for dairy cows: A review and a meta-analysis. Can. J. Anim. Sci., 91, 529–543.
Huuskonen, A., Huhtanen, P. & Joki-Tokola, E., 2014. Evaluation of protein supplementation
for growing and finishing cattle fed grass silage-based diets: A meta-analysis. Animal. (in
press).
Lammers, B.P. & Heinrichs, A.J., 2000. The response of altering the ratio of dietary protein
to energy on growth, feed efficiency, and mammary development in rapidly growing
prepubertal heifers. J. Dairy Sci., 83, 977–983.
Marini, J.C. & Van Amburgh, M.E., 2003. Nitrogen metabolism and recycling in Holstein
heifers. J. Anim. Sci., 81, 545–552.
MTT 2014. Feed tables and nutrient requirements. MTT Agrifood Research Finland.
Retrieved 8 May 2014, from http://www.mtt.fi/feedtables
Schwab, C.G., Huhtanen, P., Hunt, C.W. & Hvelplund, T., 2005. Nitrogen requirements of
cattle. In: E. Pfeffer and A. Hristov (eds.). Nitrogen and Phosphorus Nutrition of Cattle. Cabi
Publishing: Wallingford, UK. pp. 13–70.
St-Pierre, N.R., 2001. Invited review. Integrating quantitative findings from multiple studies
using mixed model methodology. J. Dairy Sci., 84, 741–755.
Proceedings of the 5th Nordic Feed Science Conference
77
Ruminant nutrition
A meta-analysis of milk production responses to increased supply of AAT
I. Schei1, N.I. Nielsen2, M. Åkerlind3 & H. Volden1,4
1TINE Norwegian Dairies, 1431 Ås, Norway; 2Knowledge Centre for Agriculture, Cattle,
8200 Aarhus N, Denmark; 3Växa Sverige, 751 05 Uppsala, Sweden; 4Norwegian University
of Life Sciences, Department of Animal and Aquacultural Sciences, P.O. Box 5003, N-1432
Ås, Norway
Correspondence: [email protected]
Introduction
Feed protein is the single most expensive nutritional factor in the feed ration of dairy cows.
To obtain a high milk production, dairy cows are dependent on high quality feed protein often
resulting in an overfeeding or imbalance between the requirement and utilization of feed
protein leading to increased excretion of nitrogen in urine and feces. Increased focus on
reducing the leakage of nutrients to the environment the last years has emphasized the
importance of feed planning and nutrient optimization to dairy cows to reduce nitrogen
excretion from urine and faces. Therefore, estimation of the expected marginal responses of
increased supply of amino acids absorbed in the small intestine (AAT) is needed in order to
improve the economic optimization of diets for dairy cows. The objectives of this work were
to develop response curves for ECM and milk protein production to increased supply of AAT
and to implement these models in the NorFor feed evaluation system (Volden, 2011).
Materials and Methods
Data were selected from international published protein production trials and Nordic protein
production trials differing in protein sources and levels of rumen degradable protein. Protein
sources were mainly rape seed meal, soybean meal and fish meal. Roughages were silages of
grass, grass-clover, corn and alfalfa. Feeding regimes were either total mixed rations (TMR)
or separate feeding and the rations were fed either ad libitum or restrictively. Concentrates in
rations varied from 28.7 to 75.9 % on a dry matter (DM) basis (mean 45.8). All diets from the
trials were calculated according to NorFor based on the information of the ingredients and
composition of the diets in order to obtain energy and nutrient supplies.
In the NorFor system, there is a minimum recommendation of 10 g PBV/kg DM (protein
balance in the rumen) when optimizing feed rations to dairy cows in order to secure high
rumen microbial activity. Therefore, treatments with PBV less than this level were deleted. If
this resulted in only one treatment left within a trial, the whole trial was deleted. Trials with
less than 3 g AAT/kg DM between the lowest and the highest AAT level were not included in
the final analysis. Breeds were Holstein, Norwegian Red and Swedish Red. All treatments
included multiparous cows or a mix of primi- and multiparous, but all cows were classified as
multiparous since splitting, based on lactation number, was not possible. The final dataset
consisted of 87 treatments from 32 trials. Of these, 30 treatments included cows less than 100
days in milk (DIM), 50 treatments were between 100 and 200 DIM and 7 treatments were
later than 200 DIM, with an overall average of 130 DIM.
Animal inputs to NorFor were breed, parity, body weight (BW; mean of first and last BW in
the trial period), DIM (mean of trial period) and activity (loose housed or tied up). Animal
characteristics, milk production including fat and protein content and feed intake of all
feedstuffs were inputs to NorFor. Nutrient content of individual feedstuffs were used if
reported and then supplemented with NorFor feed table values (Norfor.info). The ratio of
78
Proceedings of the 5th Nordic Feed Science Conference
Ruminant nutrition
metabolisable protein/net energy for milk (AAT/NEL; g/MJ) was calculated as AAT
available for milk production, i.e. total AAT intake minus requirement for maintenance,
growth and mobilizing/deposition, and then divided by 3.14 MJ x ECM (energy corrected
milk). Variation in feed values and milk production for the dataset is presented in Table 1.
Table 1 Variation of feed components and milk production in the data set.
Mean
STD
Min
Max
AAT/NEL, g/MJ
15.5
2.66
7.5
23.8
AAT, g/kg DM
93
12.1
63
121
PBV, g//kg DM
32
16.2
10
81
NEL, MJ/ kg DM
6.64
0.65
5.01
8.38
Fatty acids, g/kg DM
28
5.9
18
55
Sugar + starch, g/kg DM
277
92
109
439
ECM, kg/day
29.0
5.7
12.6
39.9
Protein production, g/day
946
202
422
1371
DIM
130
54
49
273
Dependent response variables were the average daily yield of ECM or milk protein
production calculated from milk yield and milk composition according to Sjaunja et al.
(1991). The 32 trials are plotted in Figure 1 and 2 and the lines between treatments indicate
the within study response of ECM and protein production on AAT/NEL. Data were analyzed
using the linear mixed effects model in the SAS-package (SAS Institute, version 9.2). The
independent continuous variables of AAT/NEL, and PBV, NEL, fatty acids, sugar+starch
(per kg DM) and DIM and their quadratic effects were included in the full model. However,
the quadratic effect of AAT/NEL was replaced by its logarithmic values. Breed was included
as a fixed class effect. Trial within reference was treated as a random variable.
Results and Discussion
Estimates for ECM and milk protein responses to the parameters in the model are presented
in Table 2. The model explained 98% of the variation in both ECM and protein production.
However, the effect of AAT/NEL was not very clear. The linear effect of AAT/NEL was not
significant for ECM or milk protein production. However, there was a logarithmic effect of
AAT/NEL on ECM (P<0.05) and tendency on protein production (P<0.1). Even though the
PBV level was set to a minimum of 10 g/kg DM, there was a significantly linearly and
quadratic effect on ECM and protein production (P<0.05). This might suggest that microbial
activity was still not optimal at a minimum level of 10 g/kg DM and could have been limited
by the supply of soluble protein. Energy concentration (NEL/DM) and fatty acids (FA/DM)
did not significantly affect ECM or protein production linearly, but there was some quadratic
effect on protein production (P<0.05 and P<0.1, respectively).
Proceedings of the 5th Nordic Feed Science Conference
79
Ruminant nutrition
Figure 1 Plot of treatment-responses of AAT/NEL (g/MJ NEL) on ECM (kg/d). Lines indicate the within trial
responses.
Figure 2 Plot of treatment-responses of AAT/NEL (g/MJ NEL) on ECM (kg/d). Lines indicate the within trial
responses.
Response curves for ECM and milk protein production as a function of AAT/NEL for
Swedish Red cows are presented in Figure 3. The curves for Holstein and Norwegian Red are
parallel but at different levels. Increasing the supply of AAT/NEL from 10 to 19.5 g AAT/MJ
lead to an ECM-response of 3.3 kg/d. Highest increase in ECM-responses was observed
when increasing the AAT/NEL at the lowest levels. There was a diminishing return in the
range of 18 to 19.5 g AAT/NEL, where ECM increased by only 0.2 kg/d. The response of
80
Proceedings of the 5th Nordic Feed Science Conference
Ruminant nutrition
milk protein production to the linear effect of AAT/NEL supply was not significant, but
tended to be significant for the quadratic effect. The response curve for protein production
was not that steep in the beginning as for ECM and also showed a diminishing return with
increasing supply. An increase in the supply of AAT/NEL from 10 to 25 g AAT/NEL
increased milk protein production by 141 g/d. The marginal responses of AAT/NEL are
presented in Figure 4.
Different random statements and covariance structure were tested, and the effect of
AAT/NEL and the estimates of the response curves were quite sensitive to which covariance
structure was used.
In the NorFor-system, the minimum-level of AAT/NEL for milk production recommended is
15.0 g/MJ NEL. By changing this level from 15 to 19.5 g AAT/NEL, the response curve
suggests that ECM will increase by 0.8 kg and protein production by 38 g.
Table 2 Parameter estimates for ECM (kg/d) and milk protein production (g/d) responses based on AAT/NEL
(g/MJ)
ECM
Protein production
Intercept
5.32±9.98 n.s
-402.9±363.4 n.s
AAT/NEL
-0.65±0.42 n.s
-14.5±15.4 n.s
Ln(AAT/NEL)
14.14±5.53*
391.5±210 t
PBV
0.128±0.038**
3.11±1.41*
-0.0017±0.0005***
-0.053±0.016**
NEL
n.s
n.s
NEL2
n.s
4.02±2.37 t
FA
n.s
n.s
n.s
0.064±0.031*
0.123±0.049*
5.50±1.79**
-0.0006±0.00045***
-0.025±0.006***
0.81
29.0
388.6
944.5
0.98
0.98
PBV
2
2
FA
DIM
DIM
2
RMSEa
b
AIC
2
R
***P≤0.001, **P≤0.01, *P≤0.05, t=P≤0.1, ns=P>0.1; aRoot mean square error; bAkaike’s Information Criteria
Conclusions
ECM and protein production responded to increased supply of AAT according to the law of
diminishing return with increasing supply. The maximum ECM-yield was observed at 19.5 g
AAT/MJ NEL, but maximum could not be estimated for protein production within practical
levels of AAT/NEL. This suggests that the minimum recommendation of 15.0 g/MJ in
NorFor should be increased if high milk production is important. However, the response
curves are very sensitive for the covariance structure chosen in the model, and this might
explain why the maximum yield level is estimated so high. Moreover, the costs of increased
AAT/NEL supply should also be considered.
Proceedings of the 5th Nordic Feed Science Conference
81
Ruminant nutrition
Figure 3 Response curves for ECM and protein production by AAT/NEL (g/MJ) for the breed SRB.
Figure 4 Marginal production responses of AAT/NEL (g/MJ) supply for SRB breed.
References
NorFor Feedtable, http://norfor.info/Feedtab.asp
Sjaunja L.O., Bævre, L., Junkkarinen,L., Pedersen, J., Setälä, J. 1990.A Nordic proposal for
an energy corrected milk (ECM) formula. Paper presented at the 27th Session of the
International Committee for Breeding and Productivity of Milk Animals, Paris, 2-6 July.
Volden, H., 2011. Overall model description. In: 428 Volden, H. (Ed.), NorFor - The Nordic
Feed Evaluation System. Wageningen Academic Publishers, Wageningen, the Netherlands,
pp. 23-26.
82
Proceedings of the 5th Nordic Feed Science Conference
Ruminant nutrition
Dietary nutrient density and effects on intake and production
S. Rengman, B. Johansson & M. Murphy
Lantmännen Lantbruk, Feed development, 205 03 Malmö, Sweden
Correspondence: [email protected]
Introduction
Dairy cow dry matter (DM) intake is influenced by both physical limits to rumen fill and
fulfilment of metabolic requirements. Diet composition in different systems can vary,
especially regarding the carbohydrate fraction, despite similar dietary goals, and may affect
rumen fill and metabolic satiation differently. The Norfor system calculates digestibility and
absorption in the entire digestive tract as well as calculating post absorption metabolism with
a strong focus on least cost and maximum forage intake. The LFU system is mainly focused
on digestion kinetics in the rumen and recommends a dietary nutrient composition to obtain a
biological optimum for rumen metabolism to sustain the actual milk production. While both
systems have similar approaches regarding rumen metabolism there are some significant
differences, especially regarding fiber kinetics and interactions with non-structure fibers. The
Norfor system models for interactions between substrates while the LFU system alters diet
composition according to milk yield. The objective of this study was to compare rations
composed according to Norfor and the LFU system
Materials and Methods
To compare the effects of diets composed after Lantmännen’s LFU system and the Nordic
program, Norfor, a trial was conducted at ‘Nötcenter Viken’, Falköping, Sweden with 48
cows in 3 groups in a continuous experiment. The groups were fed diets according to the
Norfor system (Norfor), according to the LFU system (LFU) and a diminished LFU diet in
which the fiber fraction was slightly improved in relation to LFU recommendations but with
less starch (LFUdim). The Norfor diet was composed by an advisor from ‘Växa’.
Table 1 Diet composition, kg dry matter, of the 3 diets offered in feeding trial
Feed
Norfor
LFU
LFUdim
Grass silage, 2 cut
12.4
10.0
10.0
Maize silage
5.0
2,4
1,5
nd
Barley, crushed
3.1
Unik 52
1.0
Unik Prestige
3.5
Solid 120
Solid 350
11,5
9.3
Solid 520
5.0
Mineral mix
0.1
0.1
0.1
Total DM
26.8
24.1
24.1
Cows (76 – 175 days in milk (DIM), first and second parity, were grouped and blocked after
breed, DIM and milk yield. Following a 2 week pre-period, the diets were fed for 3 weeks.
Average milk yield was 38 kg d-1 at blocking and diets were composed for a 40 kg production
level. At the start of the trial (Dec 16) DIM for Norfor, LFU and LFUdim were; 121, 125 and
Proceedings of the 5th Nordic Feed Science Conference
83
Ruminant nutrition
123, respectively. The trial was approved by the Ethics Committee for Animal Experiments,
Göteborg.
All diets were fed ad lib as total mixed rations (TMR). New feed was offered once a day.
Diet composition is presented in Table 1. All feeds on the farm were available for use as well
as all standard feed products sold by Lantmännen. Silages were a second harvest mixed grass
and clover ley and/or a corn silage. Both silages were grown and harvested at Nötcenter
Viken, Sweden. The grass/clover silage had a DM content of 38% with 147 g crude protein
(CP) and 446 g neutral detergent fibre (NDF) kg-1 DM. The corn silage had a DM content of
33% with 84 g CP, 415 g NDF and 258 g starch kg-1 DM. The rumen degradability of NDF
(EFD) for the grass/clover and corn silages were 0.573 and 0.450, respectively, mean of 3
analyses.
Silages were sampled weekly and analyzed at the Kungsängen laboratory, SLU, Uppsala, for
DM, protein, starch, neutral detergent fibre (NDF), fermentation acids and ash. In addition
rumen protein and fiber degradation were analyzed by in sacco by the Norfor method and
calculations for rumen degradability according to Kristensen et al. (1982).
Cows were housed in a free-stall barn and milked 3 x d-1 in a DeLaval carousel. Feed was
distributed into bins resting on load cells and the decrease in mass was recorded
automatically. Access to bins was governed by gates regulated by the cows’ transponders
(BioControl, Rakkestad, Norway) . Each group of 16 cows had access to 8 bins. Dry matter
intake (DMI) and milk yield were monitored daily. Milk composition was monitored the last
week of the trial.
Table 2 Nutrient content (g kg-1DM) of the three diets offered in the feeding experiment and the LFU
recommendations for a milk yields of 40 kg (LFU Rec)
Nutrient
Crude Protein
Norfor
LFU
LFUdim
LFU Rec.
174
186
189
180
a
119
123
121
113
b
56
63
67
67
Starch
136
190
151
190
NDF
381
326
350
340
Rumen degradable NDF
203
179
190
180
Rumen undegradable NDF
178
147
160
160
Ether extract (EE)
39
46
47
50
0.533
0.549
0.543
0.681
0.661
0.643
RDP
UDP
EFD
c
EPD
a
d
b
c
Rumen degradable protein; Rumen udegradable protein; Fraction of NDF degraded in the rumen, estimated
from individual feed values; dFraction of crude protein degraded in the rumen, estimated from individual feed
values.
Results were analysed using a one-way Anova with Minitab, version 16. Results are
presented as group averages. Statistical significance was P <0.005.
84
Proceedings of the 5th Nordic Feed Science Conference
Ruminant nutrition
DMI (kg DM) and milk yield (kg)
40
k
g
35
30
D
M 25
k 20
g
15
m
i 10
l
k
5
B
38
C
34.8
b
24.2
0
Norfor
A
36.1
a
25.6
ab
25.2
LFU
LFUdim
Figure1 DMI (kg DM d-1) and milk yield (kg d-1) for the 3 diets Norfor, LFU and LFUdim. Different letters
denote significant differences, (0.005).
Results and Discussion
LFU recommendations for a production of 40 kg assume an intake of 22.9 kg and nutrient
densities are based on that intake. DMI intake with the Norfor diet was only 24.2 kg, 2.6 kg
less than predicted (Table 1). DMI with both LFU diets was 25 kg, 2.1 kg more than
predicted (Table 1).
None of the diets had an EE content as high as recommended by the LFU system (5% of
DM), but the LFU diet supplied the recommended amount due to a higher intake (Table 2).
The LFU diet met the LFU recommended concentrations for CP and starch but was low in
NDF and 0.4% units lower than the recommended rumen undegradable dietary protein
(UDP). The Norfor diet had higher concentrations of NDF, rumen degradable NDF and
rumen undegradable NDF but was lower in CP, UDP and starch than recommended by the
LFU system (180, 67 and 190 g kg-1DM, respectively). LFUdim had more CP, rumen
degradable protein (RDP), NDF and rumen degradable NDF than recommended but less
starch.
Due to an increased DMI with the LFU diet NDF intake (8.1 kg) was higher than
recommended (7.7 kg). The NDF intake, despite a lower DMI, was even greater for the
Norfor diet (9.1 kg). Intake of rumen undegradable NDF was 3.8 kg for the LFU diet which is
only slightly higher than the recommended 3.7 kg, but was 4.5 kg for the Norfor diet, which
may have limited DMI due to rumen fill. Intake of rumen undegradable NDF was 4.1 kg with
LFUdim.
Despite a lower density than recommended, UDP intake with the LFU diet (1575 g) was
greater than recommended (1534 g) due to a higher DMI. UDP intake with LFUdim (1675 g)
Proceedings of the 5th Nordic Feed Science Conference
85
Ruminant nutrition
was much greater than recommended, which was also due to a higher DMI. Despite this, milk
protein content with the LFUdim diet was numerically lower than with the LFU diet (Table
2). This may be associated with the lower starch content and a consequent diverting of amino
acids to gluconeogenesis. UDP intake with the Norfor diet was only 1355 g, which was179 g
less than recommended.
Table 2 Milk yield, kg d-1 and milk composition, group averages, for cows fed the 3 experimental diets; Norfor,
LFU and LFUdim, in the last week of the trial
Norfor
LFU
LFUdim
Milk yield
34.2c
37.5b
36.0a
Milk fat %
4.25
4.08
4.05
Milk protein %
3.36b
3.50a
3.42ab
Milk Urea mmols/l
4.96b
5.28a
5.37a
a,b,c
Rows with different superscripts are significantly different.
Milk yield was significantly lowest with the Norfor diet. Milk fat content was numerically
highest with the Norfor diet but the difference was not statistically significant. All diets had a
higher fat content than the herd average, 3.9%. The increased fat content with Norfor is
consistent with a larger roughage fraction. Fat yield was numerically higher with the LFU
diets; 1.45, 1.53, and 1.46 kg for Norfor, LFU and LFUdim, respectively. The protein content
with the Norfor diet was lower than the herd average, 3.4. The protein content with Norfor
was numerically lowest and significantly lower than the LFU diet. Milk urea content was
significantly lower with Norfor compared to the LFU diets. The lower milk yield and lower
protein content with the Norfor diet may be associated with the lower UDP and starch
intakes.
Cows on both LFU diets were able to maintain the production level and with LFU even
increase their yields the first half of the trial (Fig. 2). Cows on the Norfor diet decreased milk
production directly after start.
Grass silage and maize silage were estimated to cost 1.55 SEK kg-1 DM. These costs were
based on a crop production and harvest cost of 1.35 SEK plus 0.20 SEK for labor and fuel in
handling the silages for feeding. The concentrates and other feeds were assigned actual
market prices. Per kg DM, the Norfor diet cost the least. Income from milk was calculated
using prices from the major dairy cooperative, ‘Arla räknesnurra’ (10 March, 2014). Even
though the feed costs were lowest for Norfor, milk minus feed costs for the entire period was
highest for the LFU diets. Feed costs and milk prices will vary over time and, therefore, an
exact relation cannot be established from a single trial. However, the results do not indicate
that higher feed costs for rations supporting a higher milk production are negative for the
producer’s economy.
Norfor had a numerically lower feed efficiency than the LFU diets. However, variation
among cows and over time was large and no differences were significant.
86
Proceedings of the 5th Nordic Feed Science Conference
Ruminant nutrition
Kg/day
Milk yield
40
39
38
37
36
35
34
33
32
31
30
1
3
5
7
9
11
13
15
17
19
Days
NorFor
LFU
LFU dim
Figure 2 Milk yield, group averages, for cows fed the 3 experimental diets
Conclusions
Norfor overestimated the intake capacity of the cows while LFU underestimated DMI. The
composition of the Norfor diet did not enable an increased DMI. Systems with the aim of
maximizing roughage intake without a decreased milk yield must have a very good estimate
of fiber digestion and passage kinetics, which apparently was not so in this trial. A high milk
production level is dependent on gluconeogenic substances which are mainly originating
from rumen propionate production and absorption of starch in the small intestines. Both of
these are closely linked to the starch content of the diet, which was much lower on the Norfor
diet, and this may also have influenced milk protein content negatively. Also the differences
in milk production between LFU and LFUdim indicate the importance of starch in the diets.
References
Kristensen, E.S., Möller, P.D. & Hvelplund, T., 1982. Estimation of the effective protein
degradability in the rumen of cows using the nylon bag technique combined with outflow
rate. Acta Agric. Scand. Suppl. 32, 123-127.
Proceedings of the 5th Nordic Feed Science Conference
87
Ruminant nutrition
Intra-ruminal mixing of concentrate pellets produced by conventional pellet press or
extruder
M. Nordqvist, P. Lund, A. C. Storm, M. R. Weisbjerg, and M. Larsen
Department of Animal Science, Aarhus University, Blichers Allé D20, DK-8830 Tjele,
Denmark
Correspondence: [email protected]
Introduction
Commercial compound concentrates used in dairy herds are for the vast majority pelleted
using a conventional pellet press. Traditionally, low rumen pH have been attributed to high
allowances of starch rich concentrates; however, recent observations have shown that pH in
ventral rumen liquid decreased when calves were fed pelleted concentrates, irrespective of
high or low starch content (Khafipour et al., 2009; Kristensen et al., 2009). This indicates that
low pH in the ventral rumen could be more linked to physical characteristics of the pellets
rather than starch content.
In the fish feed industry, extrusion has long been used to pelletize compound feeds in order to
obtain physical functional properties needed to avoid rapid sinking and disintegration of
pellets when fed in water. Extrusion is a processing technology which gives the opportunity
to obtain a variety of functional properties of feed pellets. Depending on the setting, varying
degrees of expansion of the feed can be achieved when it is forced through the matrix. This
expansion affects feed density as well as floating ability. In addition, extrusion gives different
levels of pellet durability in liquid. Pellet durability is important to avoid quick disintegration
of feed pellets in water. The effect of extrusion on nutrient digestibility in cattle have been
investigated for some feedstuffs (Benchaar et al., 1992); yet, to the authors knowledge there
have been no reports of the effect of pelletizing by extrusion on postprandial intra-ruminal
mixing of pellets.
Hypotheses: concentrate pellets fed separately during milking has a lower impact on ventral
ruminal pH if the pellets are captured by the ruminal particle phase. The objective of the
study was to compare conventional pellets with extruded pellets on their ability to be
captured by the ruminal mat. The goal was to shift the fermentation of the concentrate pellets
from the ventral rumen to the ruminal mat, thereby reducing the local acid exposure of the
ruminal wall. This is a preliminary report, presenting some of the obtained data from a larger
study.
Materials and methods
Three lactating Danish Holstein cows fitted with ruminal, duodenal and ileal cannulas were
used in a 3 × 3 Latin square design. Cows were milked twice daily at 08.00 and at 15.30 h
and were at the same time fed concentrates. The experimental concentrates were fed at the
morning milking. The experimental concentrates were pellets produced from wheat meal (2mm screen) using three different methods: conventional pelleting (sinking in liquid, high
solubility), light extrusion (floating in liquid), or heavy extrusion (sinking in liquid, low
solubility). The cows were fed a partial mixed ration (PMR) 30 min after each milking, based
on a grass-clover silage, soy bean meal, and minerals. The PMR was low in starch; thus,
starch from experimental concentrates could be used as a marker for the ability of concentrate
pellet to be captured in the particle phase of the rumen. The PMR was fed restrictedly at 95%
of ad libitum intake determined individually for the cows in a pre-experimental period. The
88
Proceedings of the 5th Nordic Feed Science Conference
Ruminant nutrition
concentrates fed at 15.30 h also had a low starch concentration and was based on dried beet
pulp, vegetable fat, and molasses.
Samples of medial rumen content as well as rumen fluid from the medial and ventral parts of
the rumen, were obtained at -0.5, +0.5, +1.5, +2.5, +4.0, +5.5, +7.0, +8.0, +9.0, +10.0, +11.5
h relative to morning feeding. Medial rumen content and fluid was sampled by hand from the
ruminal particle phase through the ruminal cannula. Two large handfuls were obtained 10 to
15 cm beneath the mat surface in the middle section of the rumen. The medial rumen particle
associated fluid, was squeezed from the samples through a single layer of cheese cloth into a
45-mL test tube. The ventral rumen fluid was sampled with a suction strainer (#RT, Bar
Diamond Inc., Parma, ID, USA) and a 60-mL syringe. Starch in medial rumen contents were
determined by enzymatic liberation of glucose bound in starch, using an YSI analyser (Dglucose oxidase, YSI 2900, YSI Inc., Yellow Springs, OH, USA) to quantify liberated
glucose.
Data was analysed using the Mixed procedure of SAS with period, processing method, time
relative to feeding, and processing × time interaction as fixed effects. Cow was considered a
random effect and time relative to feeding as repeated measurements.
Results and discussion
The starch concentration in medial rumen content increased from 0.8 ± 0.5 at -0.5 h to 4.6 ±
1.6 % of DM at +0.5 h, relative to morning feeding (Figure 1, P < 0.001). It was unaffected
by processing method of pellets when all the 11 time points were included in the dataset.
However, considering only samples obtained 5.5 h after morning feeding and later, there was
a greater starch concentration in medial rumen content for the light extrusion pellets, as
compared to conventional pellets (P = 0.015). On average across all the time points, starch
concentration in medial rumen contents were numerically greater with heavy extrusion pellets
as compared to conventional pellets (P = 0.117).
Medial rumen starch content
14
Extruding light
Extruding heavy
Conventional
Starch, % of DM
12
10
8
6
4
2
0
n=3
-0.5 0.5 1.5 2.5
4.0
5.5
7.0 8.0 9.0 10.0 11.5
Time relative to morning feeding, h
Figure 1 Starch concentration in medial rumen contents from 3 cows fed wheat meal pelletized conventionally
or by extrusion giving either light or heavy (gravimetric properties) pellets.
The ventral rumen pH (Figure 2a) in samples obtained after the afternoon milking and
feeding were not used in the analysis due to the effect on pH from feeding of concentrates.
Proceedings of the 5th Nordic Feed Science Conference
89
Ruminant nutrition
The pH in ventral rumen decreased with a similar pattern for all treatments, after feeding the
experimental concentrates, but after the nadir at +2.5 h, pH increased more rapidly with
pellets produced by extrusion (interaction, P = 0.01).
Ventral rumen pH
a)
7.2
Extruding light
Extruding heavy
Conventional
7.0
6.8
6.6
pH
6.4
6.2
6.0
5.8
5.6
5.4
5.2
n=3
-0.5
0.5
1.5
2.5
4.0
5.5
7.0
Time relative to morning feeding, h
b)
Medial rumen pH
7.2
Extruding light
Extruding heavy
Conventional
7.0
6.8
6.6
pH
6.4
6.2
6.0
5.8
5.6
5.4
5.2
n=3
-0.5
0.5
1.5
2.5
4.0
5.5
7.0
Time relative to morning feeding, h
Figure 2 Ruminal pH in liquid obtained from ventral rumen (a) and squeezed from medial grab samples (b) in 3
cows fed wheat meal pelletized conventionally or by extrusion giving either light or heavy (gravimetric
properties) pellets.
The pH in medial rumen content (Figure 2b) decreased more rapidly after feeding pellets
produced by extrusion compared with conventional pellets, but after the nadir, between 2.5
and 4 h after feeding, pH did not differ among treatments (interaction, P = 0.02). The more
rapid decrease in pH in medial rumen content with pellets produced by extrusion indicates a
greater production of acids from fermentation; thus, these pellets may have been better
captured by the particle phase. This is also supported by visual observation of intact extruded
pellets in samples obtained at +0.5 h after feeding. Observation of intact extruded pellets in
the medial mat also indicates limited crushing of pellets during eating.
90
Proceedings of the 5th Nordic Feed Science Conference
Ruminant nutrition
Conclusion
The preliminary analysis of observations indicates that wheat concentrate pellets produced by
extrusion may better mix with the rumen particle phase compared to conventional pellets.
Yet, the effects were of relatively small magnitudes indicating that the rumen system is very
robust in overcoming challenges with concentrate pellets differing greatly in physical
functional properties.
References
Benchaar, C., Vernay, M., Bayourthe, C. & Moncoulon, R., 1992. Incidence of bean (ViciaFaba) extrusion on starch and nitrogen intestinal flows in lactating cows. Repr. Nutr.
Developm. 32, 265-275.
Khafipour, E., Krause, D. O. & Plaizier, J. C., 2009. Alfalfa pellet-induced subacute ruminal
acidosis in dairy cows increases bacterial endotoxin in the rumen without causing
inflammation. J. Dairy Sci. 92, 1712-1724.
Kristensen, N.B., Storm, A.C. & Vestergaard, M., 2009. Sund drøvtyggerfodring - ved vi
hvad det er? Eksempler på fodringsmæssige muligheder for at ændre på vomfunktion hos
småkalve og slagtekalve. I: Slagtekalveproduktion - Temamøde: Ernæring, Sundhed og
Velfærd. red.: Mogens Vestergaard. Aarhus Universitet - Foulum, Det
Jordbrugsvidenskabelige Fakultet. Intern Rapport Husdyrbrug 16, 14-19.
Proceedings of the 5th Nordic Feed Science Conference
91
Ruminant nutrition
Prediction of methane emissions from dairy cows fed cold-pressed linseed cake
M. Kass1,2, A. Olt2, R. Leming2 & M. Ots1,2
1
Bio-Competence Centre of Healthy Dairy Products LLC, 1 Fr. R. Kreutzwaldi Street, 51014
Tartu, Estonia. 2 Estonian University of Life Sciences, Department of Animal Nutrition, 46
Fr. R. Kreutzwaldi Street, 51006, Tartu, Estonia
Correspondence: [email protected]
Introduction
Methane (CH4) is a normal product of anaerobic fermentation, and is hazardous to the
environment as a greenhouse gas. Methane emission is related with energy loss, 2-12% of
gross energy intake (Johnson & Johnson, 1995). Thus, in order to reduce the emission of CH4
various feed additives (e.g. oilseed crops, tannins, essential oils) have been supplemented to
the diets of dairy cattle (Meale et al., 2013), to try to avoid the reduction of production
efficiency related to CH4 emission. Since direct measurements of CH4 emission are expensive
and time-consuming there exist equations based on animal factors or nutrient intake to
determine the CH4 output by an animal (Wilkerson et al., 1995; Benchaar et al., 1998; Mills
et al., 2001; Mills et al., 2003). To reduce CH4 emission through altering rumen fermentation
parameters supplemental fat sources have been added to the diet of dairy cows (Dong et al.,
1997; Giger-Reverdin et al., 2003). Feeding crushed linseed (Beauchemin et al., 2009), crude
linseed and extruded linseed (Martin et al., 2008) to dairy cows have shown depressive
effects on ruminal methanogenesis. In addition, feeding linseed cake, which is rich in αlinolenic acid, to dairy cattle, could improve the milk fatty acid (FA) profile, which would be
beneficial for human health (Kennelly, 1996). Using mathematical models based on
numerous experiments would provide a good tool to predict the effect of diet or feed
supplement on CH4 emissions by dairy cows. In addition to dry matter intake (DMI) and diet
composition, correlations between CH4 emission and rumen volatile fatty acids (VFA)
(Ramin & Huhtanen 2013) and the milk FA composition (Chilliard et al., 2009; Dijkstra et
al., 2011) have been found. The current study used a range of prediction equations to estimate
CH4 emissions from dairy cows when they were fed cold-pressed linseed cake.
Materials and Methods
The experiment was carried out on Eerika Experimental Farm (Estonian University of Life
Sciences). Nine multiparous lactating Holstein cows were used in a 3 × 3 Latin square design,
where within each group one cow had a rumen fistula. Each experimental period lasted 28
days; comprising an adaption period of 23 days and five days of data collection. The cows
were housed individually, tethered in stalls, in which they were fed and milked. At the start of
the experiment the cows in all groups had no significant differences between them regarding
body mass, milk production or milk composition. The basal diet (on DM basis) contained
grass silage (44%), concentrate mixture (54%), and a vitamin-mineral mixture (2%). Control
group cows (no linseed) were given a concentrate mixture containing barley meal, maize
meal and soybean meal; group 2 (LL – low linseed) barley meal, maize meal, soybean meal,
heat-treated rapeseed cake and cold-pressed linseed cake (76 g/kg diet DM); and group 3 (HL
– high linseed) was fed barley meal, heat-treated rapeseed cake and cold-pressed linseed cake
(109 g/kg diet DM), respectively. All groups were fed a diet containing 11.3 MJ kg-1 DM
metabolisable energy (ME) and 180 g kg-1 DM crude protein.
92
Proceedings of the 5th Nordic Feed Science Conference
Ruminant nutrition
A list of the prediction equations based on scientific literature was chosen for CH4 estimates
(n=10, Table 1). Equations were based on either DMI, nutrient composition of the diet,
ruminal VFA or milk fatty acid (FA) composition.
Table 1 Prediction equation for CH4 emission estimates.
Equation
R2
Ellis et al., 2007
CH4 (MJ/d) = 3.23 + 0.809 × DMI (kg/d)
0.65
Eugène et al., 2008
CH4 (MJ/d) = 3.0 + 0.9 × DMI (kg/d)
Giger-Reverdin et al., 2003
CH4 (l/kg DMI) = 544.9 – 0.0220 × DMI2
Reference
DM intake
Ramin, Huhtanen, 2013
CH4 (l/d) = 20 + 35.8 × DMI − 0.50 × DMI
0.58
2
Nutrient intake
Yates et al., 2000
CH4 (MJ/day) = 1.36 + 1.21 × DM – 0.825 × DMI of
concentrate mixture + 12.8 × NDF
Giger-Reverdin et al., 2003
CH4 l/kg DMI = 47.3 – 0.0212 × DMI2 – 0.680 × %EE
0.76
Kirchgessner et al., 1995
CH4 (g/d) = 63 + 79 × CF + 10 × NFA + 26 × CP – 212 × CF
(kg/d)
0.69
Rumen volatile fatty acids
Montoya et al., 2011
CH4 (mmol/mol VFA) = 0.45 × acetate − 0.275 × propionate +
0.4×butyrate
Milk fatty acids
Chilliard et al., 2009
CH4 (g/d) = 9.46 × milk 16:0 (% of total FA) – 97.6 × milk
trans-16 +cis-14 18:1 (% of total FA) + 13.3 × forage intake (kg
of DM/d) – 78.3 × milk cis-9 14:1 (% of total FA) + 77.4 × milk
18:2n-6 (% of total FA) – 21.2
0.95
Dijkstra et al., 2011
CH4 (g/kg DM) = 24.6 + 8.74 × C17:0 anteiso − 1.97 ×trans-10
+11 C18:1 − 9.09×cis-11 C18:1 + 5.07×cis-13 C18:1
0.73
DMI – dry matter intake; EE – ether extract; CF – crude fat; CP – crude protein; NFE – nitrogen free extract;
NDF – neutral-detergent fibre.
Feed samples were analysed for chemical contents using established methods (AOAC, 2005).
Dry matter intakes were calculated based on daily intakes of diet and DM content of the
feeds. Milk samples were collected in the last five days of each experimental period. The
milk fat, protein and lactose contents were measured using infrared spectrometer MilkScan
4300 (FOSS Electric A/S, Denmark). Samples for milk FAs analyses were frozen (-18 °C)
without preservative until analysis. Milk FA profile was analysed by an Agilent 6890 gas
chromatograph equipped with a capillary column CP7420 (100 m × 0.25 mm (i.d.) with 0.25
μm film thickness). A sample of rumen fluid was collected three hours after morning feeding
from the three parts of the rumen. Samples were mixed and drained through cheesecloth, and
were taken to the laboratory for immediate VFA content analyses. The VFA content in rumen
fluid were determined chromatographically using an Agilent Technologies 7890A GC system
with a column packed with 80/120 Carbopack B-DA/4% Carbowax 20 M (Sigma-Aldrich
Corp., USA) by the method described by Faithfull (2002). Statistical analyses were
performed with MS Excel and SAS software (version 9.1; SAS Institute, 2003).
Proceedings of the 5th Nordic Feed Science Conference
93
Ruminant nutrition
Results and Discussion
Dry matter intake and ME intake were not affected by the linseed cake supplementation,
which is in agreement with a previous study (Gonthier et al., 2004). However, Martin et al.
(2008) reported a reduced DMI and gross energy intake with inclusion of extruded linseed to
the diet. Cows fed linseed cake had higher milk yields (P<0.001) and milk fat yields
(P<0.01), but lower milk fat contents (P<0.01) compared to control cows. Martin et al. (2008)
and Gonthier et al. (2005) found a decrease in milk yield when feeding linseed to dairy cows.
The total ruminal VFA and individual VFA were not affected by the treatment, with the
exception of propionate which was significantly increased (P<0.01) with linseed inclusion in
the diet. The cold-pressed linseed cake supplementation of the diet had a significant effect
(P<0.05) on the proportions of saturated FA, mono and polyunsaturated FA, and α-linolenic
acid content in milk fat. This agrees with the findings of Caroprese et al. (2010), that linseed
supplementation improves milk composition, thereby increasing polyunsaturated FA contents
in milk fat.
Using prediction equations based on DM intakes showed that as DMI was not affected by the
dietary treatment, there was no effect on CH4 emission (Table 2). This confirms the results of
Ellis et al. (2007) that methane emission is largely related to DMI. Calculation with equations
based on the nutrient component (Giger-Reverdin et al., 2003; Kirchgessner et al., 1995),
showed decreased CH4 emissions with linseed supplementation, which is in agreement with
Martin et al. (2008) that dietary supplementation based on linseed, will depress ruminal
methanogenesis.
Table 2 Effect of linseed supplement of CH4 emission predicted by different formulae
References, unit
Treatment
SE
LL
21.8a
22.1b
22.4
0.49
NS
23.7
a
24.0
b
24.3
0.54
NS
Giger-Reverdin et al., 2003, l/kg DMI
33.3
a
32.7
b
32.5
0.61
NS
Ramin, Huhtanen, 2013, l/day
578
576a
586b
8.0
NS
Yates et al., 2000, MJ/day
23.0
23.6
23.5
0.46
0.09
Giger-Reverdin et al., 2003, l/kg DMI
20.3
a
17.8
b
b
1.80
<0.001
Kirchgessner et al., 1995, g/ day
445c
366a
393b
8.1
<0.001
Montoya et al., 2011, mmol/mol VFA
312
311
299
14.3
a
c
b
9.4
<0.001
0.13
<0.001
Ellis et al., 2007, MJ/day
Eugene et al., 2008, MJ/day
Chilliard et al., 2009, g/day
535
Dijkstra et al., 2011, g/kg diet DM
26.6a
401
HL
P
Control
25.4c
17.8
450
26.1b
NS
a,b,c
– within a row a means with different superscripts are different at P<0.05; LL – low linseed diet, HL – high
linseed diet, NS – not significantly different (P>0.05).
As linseed feeding only altered rumen propionate concentration, the equation based on rumen
VFA, by Montoya et al. (2011), showed that dietary manipulation had no significant effect on
CH4 emission. The effect of linseed feeding on milk composition has been studied previously
(Glasser et al, 2008), therefore the CH4 emission affected by the dietary fat source was also
evaluated, using milk FA composition as a prediction tool. In the current study, calculations
94
Proceedings of the 5th Nordic Feed Science Conference
Ruminant nutrition
using the equation based on milk FA (Chilliard et al., 2009; Dijkstra et al., 2011) showed that
feeding linseed cake reduced CH4 production.
Conclusions
The results of this study, using different equations to predict CH4 emission in dairy cows, are
generally in agreement with previous studies. Adding cold-pressed linseed cake to the diet of
dairy cows showed that oilseed supplementation could potentially reduce CH4 emission.
Acknowledgements
The research leading to these results was financed by the European Community’s Regional
Development Fund in the framework of the Competence Centre Programme of Enterprise
Estonia under project no. EU30002 carried out by the Bio-Competence Centre of Healthy
Dairy Products.
References
AOAC, 2005. Official Methods of Analysis of AOAC INTERNATIONAL, 18th ed. –
Association of Official Analytical Chemists International, Gaithersburg, MD, USA.
Beauchemin, K.A., McGinn, S.M., Benchaar, C. & Holtshausen, L., 2009. Crushed
sunflower, flax, or canola seeds in lactating dairy cow diets: Effects on methane production,
rumen fermentation, and milk production. J. Dairy Sci., 92, 2118–2127.
Benchaar, C., Rivest, J., Pomar, C. & Chiquette, J., 1998. Prediction of methane production
from dairy cows using existing mechanistic models and regression equations. J. Anim. Sci.,
76, 617–627.
Caroprese, M., Marzano, A., Marino, R., Gliatta, G., Muscio, A. & Sevi, A., 2010. Flaxseed
supplementation improves fatty acid profile of cow milk. J. Dairy Sci., 93, 2580–2588.
Chilliard, Y., Martin, C., Rouel, J. & Doreau, M., 2009. Milk fatty acids in dairy cows fed
whole crude linseed, extruded linseed, or linseed oil, and their relationship with methane
output. J. Dairy Sci., 92, 5199–5211.
Dijkstra, J., van Zijderveld, S.M., Apajalahti, J.A., Bannink, A., Gerrits, W.J.J., Newbold,
J.R., Perdok, H.B. & Berends, H., 2011. Relationships between methane production and milk
fatty acid profiles in dairy cattle. Anim. Feed Sci. Technol., 166–167, 590–595.
Dong, Y., Bae, H.D., McAllister, T.A., Mathison, G.W. & Cheng, K-J., 1997. Lipid-induced
depression in methane production and digestibility in the artificial rumen system (RUSITEC).
Can. J. Anim. Sci., 77, 269–278.
Ellis, J.L., Kebreab, E., Odongo, N.E., McBride, B.W., Okine, E.K. & France, J., 2007.
Prediction of methane production from dairy and beef cattle. J. Dairy Sci., 90, 3456–3467.
Eugène, M., Massé, D., Chiquette, J. & Benchaar, C., 2008. Meta-analysis on the effects of
lipid supplementation on methane production of lactating dairy cows. Can. J. Anim. Sci., 88,
331–337.
Faithfull, N.T., 2002. Methods in Agricultural Chemical Analysis: a Practical handbook,
CABI Publishing, UK, 266 pp.
Giger-Reverdin, S., Morand-Fehr, P. & Tran, G., 2003. Literature survey of the influence of
dietary fat composition on methane production in dairy cattle. Livest. Prod. Sci., 82, 73–79.
Proceedings of the 5th Nordic Feed Science Conference
95
Ruminant nutrition
Glasser, F., Ferlay, A. & Chilliard, Y., 2008. Oilseed lipid supplements and fatty acid
composition of cow milk: A meta-analysis. J. Dairy Sci., 91, 4687–4703.
Gonthier C., Mustafa, A.F., Berthiaume, R., Petit, H.V., Martineau, R. & Ouellet, D.R., 2004.
Effects of feeding micronized and extruded flaxseed on ruminal fermentation and nutrient
utilization by dairy cows. J. Dairy Sci., 87, 1854–1863.
Gonthier C., Mustafa, A.F., Ouellet, D.R., Chouinard, P.Y., R Berthiaume, R. & Petit, H.V.,
2005. Feeding micronized and extruded flaxseed to dairy cows: Effects on blood parameters
and milk fatty acid composition. J. Dairy Sci., 88, 748–756.
Johnson, K.A., Johnson, D.E., 1995. Methane emissions from cattle. J. Anim. Sci., 73, 2483–
2492.
Kennelly, J.J., 1996. The fatty acid composition of milk fat as influenced by feeding
oilseeds. Anim. Feed Sci. Technol., 60, 137–152.
Kirchgessner M., Windisch W. & Müller H.L., 1995. Nutritional factors for the quantification
of methane production. In: Ruminant physiology: Digestion, metabolism, growth and
reproduction, Proceedings of the 8th International Symposium on Ruminant Physiology (eds.
Engelhardt, Leonhard-Marek, Breves & Giesecke), Ferdinand Enke Verlag, Stuttgart, 1995,
pp. 333–348.
Martin, C., Rouel, J., Jouany, J.P., Doreau, M. & Chilliard, Y., 2008. Methane output and diet
digestibility in response to feeding dairy cows crude linseed, extruded linseed, or linseed oil.
J. Dairy Sci., 86, 2642–2650.
Meale, S.J., McAllister, T.A., Beauchemin, K.A., Harstad, O.M. & Chaves, A.V., 2013.
Strategies to reduce greenhouse gases from livestock. Acta Agric Scand., Section A – Anim.
Sci., 62, 199–211.
Mills, J., Dijkstra, J., Bannink, A., Caunnell, S.B., Kebreab, E. & France, J., 2001. A
mechanistic model of whole-tract digestion and methanogenesis in the lactating dairy cow:
Model development, evaluation, and application. J. Anim. Sci., 79, 1584–1597.
Mills, J., Kebreab, E., Yates, C.M., Crompton, L.A., Caunnell, S.B., Dhanoa, M.S., Agnew,
R.E. & France, J., 2003. Alternative approaches to predicting methane emissions from dairy
cows. J. Anim. Sci., 81, 3141–3150.
Montoya, J.C., Bhagwat, A., Peiren, N., De Campeneere, S., De Baets, B. & Fievez, V.,
2011. Relationships between odd- and branched-chain fatty acid profiles in milk and
calculated Enteri methane proportion for lactating dairy cattle. Anim. Feed Sci. Technol.,
166, 596–602.
Ramin, M. & Huhtanen, P., 2013. Development of equations for predicting methane
emissions from ruminants. J. Dairy Sci., 96, 2476–2493.
SAS Institute, 2003. SAS Online DOC, version 9.1. SAS Institute Inc, Cary, NC.
Wilkerson, V.A., Casper, D.P.& Mertens, D.R., 1995. The prediction of methane production
of Holstein cows by several equations. J. Dairy Sci., 78:2402–2414.
Yates, C.M., Cammell, S.B., France, J. & Beever, D.E., 2000. Prediction of methane
emissions from dairy cows using multiple regression analysis. Proc. Br. Soc. Anim. Sci., 94.
96
Proceedings of the 5th Nordic Feed Science Conference
Ruminant nutrition
Effects of concentrate feeding strategies on performance of dairy bulls
K. Manni1,2, M. Rinne2 & A. Huuskonen3
1
HAMK University of Applied Sciences, 31310 Mustiala, Finland, 2MTT Agrifood Research
Finland, Animal Production Research, 31600 Jokioinen, Finland, 3MTT Agrifood Research
Finland, Animal Production Research, 71750 Maaninka, Finland,
Correspondence: [email protected]
Introduction
The main factor that affects the profitability of beef production is feeding costs and animal
performance. The concentrate feeding strategy may affect substantially animal performance
(Keane et al., 2006; Huuskonen et al., 2007) and thus the efficiency of beef production. High
levels of concentrates are typically used in intensive beef production to achieve high growth
rate and carcass conformation. However, good quality silage alone can support high level of
performance of growing cattle (Randby et al., 2010). There is also a possibility to divide the
growth into periods by restricting feed intake during part of the growing period. In that case
growth rate of animals can be manipulated, and compensatory growth can be induced after
feed restriction with possible benefits in feed efficiency (Hornick, et al., 2000). In this work
the effects of concentrate feeding strategies on performance of dairy bulls was studied.
Materials and methods
The feeding experiment comprised in total 36 dairy bulls, 20 Holstein-Friesian and 16 Nordic
Red. The bulls were taken into the experiment at an average live weight (LW) of 230 (s.d.
±36.9) kg and an average age of 200 (±24.9) days. In the beginning of the experiment, the
bulls were divided into nine blocks of four animals each by LW and breed so that there were
five Holstein-Friesian blocks and four Nordic Red blocks. Within the block, the bulls were
randomly allotted to one of the four treatments. During the experiment, three animals were
eliminated due to reasons unrelated to the treatments.
During the feeding experiment the animals were housed in a tie stall barn. They were fed ad
libitum (proportionate refusals of 5%) either grass silage alone or a total mixed ration (TMR),
individually. The silages were prepared from primary growth of timothy stands, harvested at
early heading stage of timothy using a mower conditioner, wilted for 5 h and harvested using
a precision-chop forage harvester. Silages were ensiled in bunker silos and treated with a
formic acid based additive applied at a rate of 5 l/tonne of fresh forage. The concentrate used
was rolled barley. The feeds were analysed as described by Huuskonen (2013). Silage
fermentation quality was analysed by electrometric titration as described by Moisio and
Heikonen (1989). Concentrations of metabolizable energy (ME), amino acids absorbed from
the small intestine (AAT) and protein balance in the rumen (PBV) values were calculated
according to the Finnish feed evaluation system (MTT 2014). The whole experimental
period, 12 months, was divided into two parts, early and late, both lasting for 6 months. The
four dietary treatments were:
1. GS: Grass silage alone during the whole experimental period
2. SC: Steady concentrate allowance. TMR contained grass silage (700 g/kg dry matter
(DM)) and barley (300 g/kg DM) during the whole experimental period.
3. IC: Increased concentrate allowance. Grass silage alone during the early part of the
experiment and then TMR containing grass silage (400 g/kg DM) and barley (600 g/kg DM)
during the late part of the experiment.
Proceedings of the 5th Nordic Feed Science Conference
97
Ruminant nutrition
4. DC: Decreased concentrate allowance. TMR contained grass silage (400 g/kg DM) and
barley (600 g/kg DM) during the early part of the experiment and then grass silage alone
during the late part of the experiment.
The target average slaughter age was 560–570 days. After slaughter the carcasses were
weighed hot and classified for conformation and fatness using the EUROP quality
classification (EC 2006). The cold carcass weight was estimated as 0.98 of the hot carcass
weight. Dressing proportions were calculated from the ratio of cold carcass weight to final
LW. Carcass gain was calculated as the difference between the final cold carcass weight and
the carcass weight in the beginning of the experiment divided by the number of growing
days. Carcass weight in the beginning of the experiment was assumed to be 0.50 × initial
LW.
The results are shown as least squares means. The normality of analysed variables was
checked using graphical methods: box-plot and scatter plot of residuals and fitted values. The
data was subjected to analysis of variance using the SAS MIXED procedure (version 9.1,
SAS Institute Inc., Cary, NC). Differences between the dietary treatments were tested using
three orthogonal contrasts: 1. GS vs. others, 2. SC vs. IC + DC, and 3. IC vs. DC. Used limit
for statistical significant differences was P<0.05 and P<0.1 was limit for tendency.
Results
The grass silage quality was good both in terms of feed values and preservation quality (pH
3.97, DM 234 g/kg, 161 g crude protein (CP), 556 g neutral detergent fibre (NDF), 55 g lactic
and formic acid, 12 g volatile fatty acids, 50 g water soluble carbohydrates and 11.3 MJ ME
per kg DM, and 52 g ammonia N per kg N. The barley had a typical chemical composition
and feed values of 13.2 MJ ME and 120 g CP per kg DM.
The results are presented in Tables 1 and 2. Including concentrate in the diet increased total
DM intake (DMI) during the whole experimental period from 7.97 to 8.55 kg/d (P<0.01)
when GS was compared to other feeding groups, and subsequently increased ME (P<0.001),
AAT (P<0.001) and starch (P<0.001) intake. Concentrate intake decreased NDF intake
(P<0.001) and PBV (P<0.001). Removing concentrate from the diet either in the early (IC) or
late (DC) part of the growing period, decreased DMI (P<0.001). The DMI during the total
experimental period, was greater in IC compared to DC (P<0.05).
Live weight gain (LWG) of the bulls consuming silage alone in the diet was as high as 1119
g/d when the whole growing period was observed. Including concentrate into the diet (SC, IC
and DC), increased daily growth rate further to 1222 g/d (P<0.1). The average concentrate
intake (SC, IC and DC) during the whole growing period was 2.70 kg DM/d and growth
response to 1 kg DM increased concentrate intake was 38 g/d.
Concentrate intake during the early part of growing period (SC and DC) increased growth
rate 223 g/d numerically compared to bulls with silage as the sole feed (GS and IC). The
daily amount of concentrate, 300 vs. 600 g /kg DM, did not differ numerically in growth
rates.
Concentrate intake during only the late part of growing period (IC) increased daily growth
rate numerically 332 g compared to the growth rate in early part without concentrate. In the
other feeding groups (GS, SC and DC) during the late part of growing period growth rates
decreased numerically compared with the growth during the early part. When IC was
98
Proceedings of the 5th Nordic Feed Science Conference
Ruminant nutrition
compared to DC, concentrate inclusion during the late part of growing period increased total
daily growth rate by 157 g/d (P<0.05).
Concentrate inclusion increased carcass weight by 25 kg (P<0.05) and dressing proportion by
12 g/kg (P<0.05), comparing GS with other feeding groups. Conformation score tended to
increase (P<0.1) by concentrate inclusion but fat score was not affected, comparing GS with
others. Comparing IC to DC, concentrate inclusion in the late part of growing period
increased carcass weight (P<0.05) and fat score (P<0.01).
Concentrate inclusion improved feed conversion rate. Both DMI (P<0.1, P<0.05), ME
(P<0.05, P<0.05) and CP consumption (P<0.1, P<0.05) per kg LWG and kg carcass gain
decreased, respectively.
Discussion
It is well established that increasing concentrate allowance increases DM and ME intakes and
growth rates (Steen and Kilpatrick, 2000). However, when using good quality silage (high
digestibility and preservation quality), the growth response to increased concentrate intake
generally declines because the differences in intake remain marginal (Steen et al., 2002). In
the present experiment the grass silage was of high quality and bulls achieved high growth
rate, 1119 g/d, when the silage was given as a sole feed. It seems that the young animals
could not make use of the higher level of concentrate probably due to the good quality silage
because the highest level of concentrate did not seem to increase the growth rate.
The increase in growth per 1 kg DM increase in concentrate intake is consistent with results
of Huuskonen et al. (2007; 27 g/d) but there are also experiments where responses have been
greater (Manni et al., 2013; 73 g/d, Martinsson et al., 1990; 84 g/d). In IC during
compensatory growth, increase in growth per 1 kg DM increased concentrate intake was 53 g.
Increased feed conversion rate (kg DM/kg LW and MJ ME/kg LW) by concentrate inclusion
is consistent with results reported by Huuskonen et al. (2007).
When concentrate was included to the diet during the late part of growing period (IC), growth
rate increased substantially compared to the period without concentrate or to the other groups
during the late part of growing period (Figure 1). The faster growth rate can be attributed
partly to the compensatory growth potential and partly to the higher concentrate proportion in
the diet at that time. The animals in IC had capacity to catch up the growth during the late
part of growing period that their weight at the end of the experiment was even slightly higher
compared to other groups. This indicates that the bulls made use of the concentrate intake
during the late part of growing period.
Concentrate inclusion increased carcass weight, dressing proportion and tended to improve
carcass conformation which is consistent with many previous experiments (Keane et al.,
2006, Keane and Fallon, 2001) but in contrast to Manni et al. (2013) and Huuskonen et al.
(2007). Concentrate intake did not affect the carcass fat score, consistent with Randby et al.
(2010) but in contrary to many other experiments such as Huuskonen et al. (2007) and Keane
and Fallon (2001). Concentrate intake during the late period increased carcass weight, but
also carcass fatness, although the carcass fat scores were in general low.
Usefulness of periodic growth rate depends on the conditions of the farm, the targets of
production and the prices of feeds used.
Proceedings of the 5th Nordic Feed Science Conference
99
Ruminant nutrition
Table 1 Feed, energy and nutrient intake of growing dairy bulls using different concentrate feeding strategies
GS
SC
IC
DC
8
7
9
9
7.09
5.30
6.84
3.08
0.121
<0.001
0.024
<0.001
7.09
7.56
6.84
7.69
0.154
0.108
0.111
<0.001
236
218
225
202
6.6
0.007
0.518
0.009
Silage, kg/day
8.84
6.51
4.64
8.91
0.192
<0.001
0.534
<0.001
Kg/day
8.84
9.31
10.93
8.91
0.270
0.005
0.064
<0.001
g/kg0.60 live weight
221
203
216
212
10.0
0.306
0.373
0.757
Silage, kg/day
7.97
5.91
5.74
5.88
0.141
<0.001
0.563
0.411
Kg/day
7.97
8.44
8.90
8.30
0.194
0.010
0.487
0.020
live weight
227
213
213
211
6.2
0.039
0.956
0.810
Nutrient intake (g/d)
Metabolizable energy, MJ/d
Crude protein
89.8
1277
99.9
1249
106.2
1274
98.3
1247
2.29
28.5
<0.001
0.514
0.396
0.734
0.010
0.465
Neutral detergent fibre
4425
3798
3854
3740
87.7
<0.001
0.990
0.305
Starch
69
1484
1838
1419
37.5
<0.001
0.003
<0.001
AAT (g/d)
681
750
792
741
17.1
<0.001
0.427
0.024
PBV (g/d)
263
130
90
143
3.6
<0.001
0.005
<0.001
Number of observations
Dry matter intake
Early part
Silage, kg/day
Kg/day
0.60
g/kg
live weight
SEM
Contrasts (P-value)
1
2
3
Late part
Total experimental period
0.60
g/kg
GS = Grass silage alone; SC = Steady concentrate, 300 g/kg dry matter, allowance; IC = Increased concentrate
allowance (first no concentrate, then concentrate 600 g/kg dry matter); DC = Decreased concentrate allowance
(first concentrate 600 g/kg dry matter, then no concentrate); SEM = Standard error of the mean; Contrasts: 1 =
GS vs. others; 2 = SC vs. IC + DC; 3 = IC vs. DC; AAT = Amino acids absorbed in the small intestine; PBV =
Protein balance in the rumen.
Conclusions
Although good quality silage as the sole feed may support high levels of performance of
growing cattle, including concentrate into the diet may improve performance of animals.
Periodic concentrate allocation demonstrates the ability of growing bulls to adapt to different
feeding regimes without affecting their performance.
100
Proceedings of the 5th Nordic Feed Science Conference
Ruminant nutrition
Table 2 Growth performance, carcass characteristics and feed conversion of growing dairy bulls using different
concentrate feeding strategies
GS
SC
IC
DC
SEM
Contrasts (P-value)
1
Number of observations
Live weight gain (LWG; g/d)
Early part
Late part
Total experimental period
Carcass gain, g/d
Carcass weight, kg
Dressing proportion, g/kg
Carcass qualification
Conformation score, EUROP
Fat score, EUROP
Feed conversion
Kg DM/kg LWG
Kg DM/kg carcass gain
MJ ME/kg LWG
MJ ME/kg carcass gain
g crude protein / kg LWG
g crude protein / kg carcass gain
2
3
8
7
9
9
1153
1086
1119
580
324
504
1378
1009
1194
642
341
518
1149
1481
1315
697
366
514
1370
945
1158
621
339
516
71.3
63.1
52.9
29.3
8.6
5.4
0.068
0.388
0.082
0.028
0.013
0.050
0.172
0.011
0.496
0.618
0.281
0.583
0.019
<0.001
0.025
0.048
0.018
0.735
4.5
2.7
5.1
3.0
5.2
3.2
4.7
2.5
0.24
0.19
0.068
0.382
0.630
0.639
0.169
0.008
7.82
15.1
93
179
1161
2243
7.12
13.2
84
155
1069
1977
6.58
12.4
77
145
996
1881
7.13
13.4
83
156
1069
2010
0.426
0.81
5.2
9.7
61.5
122.2
0.064
0.022
0.047
0.014
0.090
0.038
0.609
0.789
0.558
0.716
0.622
0.831
0.306
0.350
0.297
0.335
0.349
0.404
GS = Grass silage alone; SC = Steady concentrate, 300 g/kg dry matter allowance; IC = Increased concentrate
allowance (first no concentrate, then concentrate 600 g/kg dry matter); DC = Decreased concentrate allowance
(first concentrate 600 g/kg dry matter, then no concentrate); SEM = Standard error of the mean; Contrasts: 1 =
GS vs. others; 2 = SC vs. IC + DC; 3 = IC vs. DC; Carcass conformation score: 1 = poor, 15 = excellent;
Carcass fat score: 1 = low, 5 = very high.
Figure 1 Growth rates of dairy bulls during early (1) and late (2) part of growing period given either silage
alone or concentrate allowance being steady, increased or decreased.
Proceedings of the 5th Nordic Feed Science Conference
101
Ruminant nutrition
References
EC, 2006. Council Regulation (EC) No 1183/2006 of 24 July 2006 concerning the
Community scale for the classification of carcasses of adult bovine animals. Off. J. Europ.
Union L 214, 1–6.
Hornick, J.L., Van Eenaeme, C., Gérard, O., Dufrasne, I. & Istasse, L., 2000. Mechanisms of
reduced and compensatory growth. 2000. Domest. Anim. Endocrinol. 19, 121-132.
Huuskonen, A., 2013. Performance of growing and finishing dairy bulls offered diets based
on whole-crop barley silage with or without protein supplementation relative to a grass
silage-based diet. Agric Food Sci., 22, 424-434.
Huuskonen, A., Khalili, H., Joki-Tokola, E., 2007. Effects of three different concentrate
proportions and rapeseed meal supplement to grass silage on animal performance of dairybreed bulls with TMR feeding. Livest. Sci. 110, 154-165.
Keane, M.G., Drennan, M.J. & Moloney A.P., 2006. Comparison of supplementary
concentrate levels with grass silage, separate or total mixed ration feeding, and duration of
finishing in beef steers. Livest. Sci. 103, 169–180.
Keane, M.G. & Fallon, R.J., 2001. Effects of feeding level and duration on finishing
performance and slaughter traits of Holstein-Friesian young bulls. Ir. J. Agric. Food Res. 40,
145-160.
Manni, K., Rinne, M. & Huhtanen, P., 2013. Comparison of concentrate feeding strategies for
growing dairy bulls. Livest. Sci. 152, 21-30.
Martinsson, K., 1990. The effects of forage digestibility and concentrate supplementation on
performance of finishing bulls. Swed. J. Agric. Res. 20, 161-167.
Moisio, T. & Heikonen, M., 1989. A titration method for silage assessment. Anim. Feed Sci.
Technol. 22, 341–353.
Randby, Å.T., Nørgaard, P. & Weisbjerg, M.R., 2010. Effect of increasing plant maturity in
timothy-dominated grass silage on the performance of growing/finishing Norwegian Red
bulls. Grass Forage Sci. 65, 273-286.
Steen, R.W.J. & Kilpatrick, D.J., 2000. The effects of the ratio of grass silage to concentrates
in the diet and restricted dry matter intake on the performance and carcass composition of
beef cattle. Livest. Prod. Sci. 62, 181-192.
Steen, R.W.J., Kilpatrick, D.J. & Porter, M.G., 2002. Effects of the proportions of high or
medium digestibility grass silage and concentrates in the diet of beef cattle on liveweight
gain, carcass composition and fatty acid composition of muscle. Grass Forage Sci. 57, 279291.
102
Proceedings of the 5th Nordic Feed Science Conference
Ruminant nutrition
Effects of supplementary concentrate level and separate or total mixed ration feeding on
performance of growing dairy bulls
M. Pesonen, E. Joki-Tokola & A. Huuskonen
MTT Agrifood Research Finland, Animal Production Research, Tutkimusasemantie 15, FI92400 Ruukki, Finland.
Correspondence: [email protected]
Introduction
Beef production in Finland is based mainly on raising dairy bulls born on dairy farms. In the
past, bulls fed on forage plus concentrates were generally offered their concentrate allowance
once or twice daily separately from the forage. Recently, many beef producers have changed
to using total mixed rations (TMR) to save labour and mechanise feeding. The rationale for
TMR feeding is to achieve a relatively stable rumen pH and fermentation pattern throughout
the day which would facilitate better cellulose digestion and a higher lipogenic to nonlipogenic volatile fatty acid ratio (Kaufmann 1976). Relative to dairy cows, there are only
few reports in the literature comparing TMR and separate feedings in growing and finishing
cattle. Nevertheless, Caplis et al. (2005) concluded that TMR-feeding had no effect on
growth performance or carcass traits compared with separate feeding in finishing crossbred
steers even though TMR-feeding increased silage and total dry matter (DM) intake. Keane et
al. (2006) reported that feeding a TMR increased intake at the low (375 g/kg DM) but not at a
high (750 g/kg DM) concentrate level compared to separate feeding in beef steers but feeding
method had no effects on overall live weight gain (LWG) or slaughter traits. The interest of
the present study is to obtain information concerning animal performance when growing
dairy bulls are fed either TMR or separate diet with two different concentrate levels.
Materials and Methods
A feeding experiment was conducted in the experimental barn of MTT Agrifood Research
Finland in Ruukki, Finland. A 2×2 factorial design was used to study the effects on animal
performance of (1) the increasing concentrate level, and (2) feeding grass silage and
concentrates separately or as a total mixed ration. A feeding experiment comprised in total 32
Nordic Red bulls. At the beginning of the feeding experiment the bulls with average live
weight (LW) of 145 kg were divided into eight blocks of four animals by LW. Within the
block, the bulls were randomly allotted to one of the four feeding treatments (8 bulls per
treatment). The four feeding treatments were: (1) grass silage (660 g/kg dry matter) plus
medium level of rolled barley (330) offered separately, (2) grass silage (660) plus medium
level of rolled barley (330) offered as TMR, (3) grass silage (330) plus high level of rolled
barley (660) offered separately, and (4) grass silage (330) plus high level of rolled barley
(660) offered as TMR. During the feeding experiment (398 days) the bulls were fed ad
libitum (proportionate refusals of 5%) either grass silage or TMR. The daily ration for all
bulls included also 150 g of a mineral mixture. A vitamin mixture was given 50 g/animal
weekly. The bulls were placed in an insulated barn in adjacent tie-stalls and were individually
fed three times per day (at 0800, 1200 and 1800 hours). Refused feed was collected and
weighed at 0700 daily. One bull (separate feeding, high level of rolled barley) was excluded
from the study due to hoof problems.
The grass silage used in the experiment was from mixed timothy (Phleum pratense) and
meadow fescue (Festuca pratensis) stand and was cut using a mower conditioner, wilted for
5 h, and harvested using a precision-chop forage harvester. The grass silage was ensiled in
Proceedings of the 5th Nordic Feed Science Conference
103
Ruminant nutrition
bunker silos and treated with a formic acid-based additive (760 g formic acid/kg, 55 g
ammonium formate/kg) applied at a rate of 5 litres/tonne of fresh grass. During the feeding
experiment silage sub-samples were taken twice a week, pooled over periods of four weeks
and stored at –20 ºC prior to analyses. The samples were analysed for DM, ash, crude protein
(CP), neutral detergent fibre (NDF), ether extracts, silage fermentation quality [pH, watersoluble carbohydrates (WSC), lactic and formic acids, volatile fatty acids] and digestible
organic matter (DOM) in DM (D-value) as described by Huuskonen (2013). Concentrate subsamples were collected weekly, pooled over periods of 12 weeks and analysed for DM, ash,
CP, NDF and ether extracts. The metabolisable energy (ME) concentration of the silages was
calculated as 0.016 × D-value. The ME concentrations of the concentrate feeds were
calculated based on concentrations of digestible crude fibre, CP, crude fat and nitrogen-free
extract described by MTT (2014). The digestibility coefficients of the concentrates were
taken from the Finnish Feed Tables (MTT 2014). Amino acids absorbed from small intestine
(AAT) and protein balance in the rumen (PBV) values were calculated according to the
Finnish feed protein evaluation system (MTT 2014).
The bulls were weighed on two consecutive days at the beginning of the experiment and
thereafter single weightings were conducted at approximately every 28 days. Before
slaughter, the bulls were weighed on two consecutive days. The target for average carcass
weight was 300 kg. Slaughter time determined based on LW and assumed dressing
proportion (0.52 g/kg) which was assessed based on earlier studies with Nordic Red bulls
(Huuskonen et al., 2007; Manni et al., 2013). The bulls were slaughtered in the Atria Ltd.
commercial slaughterhouse in Kuopio, Finland. After slaughter the carcasses were weighed
hot. The cold carcass weight was estimated as 0.98 of the hot carcass weight. Dressing
proportions were calculated from the ratio of cold carcass weight to final LW. The carcasses
were classified for conformation and fatness using the EUROP quality classification (EC
2006). The study was set up according to a complete randomized block design with animal as
an experimental unit. The results are shown as least squares means. The data were subjected
to analysis of variance using the SAS MIXED procedure (version 9.3, SAS Institute Inc.,
Cary, NC). The statistical model used was
yijkl = µ + Bk + Fi + Cj + FCij + eijkl
where µ is the overall mean, Bk is blocking effect (k=1,…,8) and eijkl is the random error
term. Fi (i=1,2), Cj (j=1,2) and FCij are the effects of feeding method (TMR, separate),
concentrate level (330, 660 g/kg DM) and their interaction, respectively.
Results and Discussion
The chemical compositions and nutritional values of the experimental feeds and total mixed
rations used in the present experiment are given in Table 1.The grass silage used was of
reasonably good nutritional quality as indicated by the D value as well as the AAT and CP
contents. The fermentation quality of the grass silage was also good as indicated by the pH
value and the low concentration of ammonia N and total acids (Table 1). Barley grain used in
the experiment had typical chemical composition and feed values. Because of the higher
energy content of barley grain compared to the grass silage, increasing the concentrate
proportion increased the calculated energy value of the ration (Table 2). Increasing the
proportion of the concentrate also increased the DM content, but decreased the NDF content
of the ration.
104
Proceedings of the 5th Nordic Feed Science Conference
Ruminant nutrition
Table 1 Chemical composition and feeding values of the grass silage, barley grain and total mixed rations
(TMR) used in the feeding experiment.
Grass silage
242
919
135
538
46
10.7
79
16
669
Barley grain
917
975
120
205
22
12.9
97
-25
NDe
TMR 330a
321
937
130
427
38
11.4
85
2
TMR 660b
476
955
125
316
30
12.2
91
-11
Dry matter (DM), g/kg feed
Organic matter (OM), g/kg DM
Crude protein, g/kg DM
Neutral detergent fibre (NDF), g/kg DM
Ether extract, g/kg DM
Metabolisable energy, MJ/kg DM
AATc, g/kg DM
PBVd, g/kg DM
Digestible OM in DM, g/kg DM
Fermentation quality of the grass silage
pH
3.72
Volatile fatty acids, g/kg DM
18
Lactic + formic acid, g/kg DM
47
In total N, g/kg
NH4-N
39
Soluble N
420
a
TMR 330 = concentrate level 330 g/kg dry matter. b TMR 660 = concentrate level 660 g/kg dry matter. c AAT
= Amino acids absorbed from small intestine. d PBV = Protein balance in the rumen. e ND = Not determined.
There were no significant interactions for intake parameters between feeding method and
concentrate level (Table 2). TMR-feeding increased total DM intake (DMI), energy intake
and CP intake of the bulls by 10, 9 and 10%, respectively, compared to separate feeding
(P<0.001). Concentrate level had no significant effects on DM or CP intake but the
increasing concentrate allowance increased energy intake 11% (P<0.01) (Table 2).
A positive effect of TMR-feeding on DMI was also noted previously in finishing steers
(Caplis et al., 2005; Keane et al., 2006) and heifers (Cooke et al., 2004). Earlier, Petchey and
Broadbent (1980) compared separate or mixed feeding of silage and concentrates for
finishing Friesian steers at silage:concentrate ratios ranging from 0 to 1.0. Mixing increased
DMI 9% with no evidence of an interaction between feeding method and concentrate level
(Petchey and Broadbent, 1980). It is suggested that some intake increases due to TMR
feeding can be explained by the rejection of unpalatable feeds in unmixed rations, something
which is not possible in TMR-feeding (Phipps et al., 1984). In another case, DMI increases
could be due to the extra processing that occurs during the mixing of the TMR or to the fact
that the whole diet is constantly available (Caplis et al., 2005).
It is well established that increasing concentrate allowance decreases silage intake and
subsequently increases ME intake (e.g. Randby et al. 2010; Manni et al. 2013) as in the
present study. In the present experiment the concentrate proportion had no significant effects
on the total DMI which is in line with results by Huuskonen et al. (2007) with dairy bulls fed
grass silage-barley-based rations. However, in many feeding experiments increasing
concentrate level has also increased total DMI to some extent (Randby et al. 2010; Manni et
al. 2013). The substitution rate (SR, decrease in silage DMI /kg increase of concentrate DMI)
in the current experiment was 0.94 and 0.84 for TMR and separate feedings, respectively.
This is in line with SR’s observed at grass silage-based feedings reported by Keane (2010)
with crossbred steers (0.82), Randby et al. (2010) with dairy bulls (0.75) and Manni et al.
(2013) with dairy bulls (0.81).
Proceedings of the 5th Nordic Feed Science Conference
105
Ruminant nutrition
Table 2 Intake, growth performance and carcass characteristics of the bulls fed either separate or total mixed
ration (TMR) feeding with two different concentrate levels.
SEMa
F
P-value
C
F×C
7
398
145
2.6
0.86
0.82
0.98
6.93
206
80.2
7.33
211
90.5
0.178
3.3
2.12
<0.001
<0.001
<0.001
0.11
0.30
<0.001
0.48
0.64
0.62
988
896
912
22.8
<0.001
0.94
0.52
576
1083
580
590
1117
609
557
1036
522
597
1136
631
13.8
33.1
19.6
0.56
0.56
0.22
0.052
0.045
0.002
0.33
0.30
0.042
7.19
13.44
82.3
153.8
7.14
13.09
87.5
160.3
6.73
13.46
77.9
155.9
6.47
11.64
79.9
143.7
0.138
0.494
1.63
5.85
<0.001
0.19
0.048
0.24
0.27
0.034
0.026
0.66
0.43
0.13
0.31
0.10
Feeding method (F)
Concentrate level (C) (g/kg DM)
TMR
330
660
Separate
330
660
Number of observations
Duration of the experiment, d
Initial live weight, kg
8
398
145
8
398
146
8
398
145
7.78
228
89.0
7.95
229
97.3
1000
Intake
Dry matter (DM), kg/d
DM intake, g/kg0.60 live weight
Metabolisable energy (ME),
MJ/d
Crude protein, g/d
Final live weight, kg
Live weight gain, g/d
Carcass gain, g/d
Feed conversion
Kg dry matter/kg LWG
Kg dry matter/kg carcass gain
MJ ME/kg LWG
MJ ME/kg carcass gain
Carcass characteristics
Carcass weight, kg
303
315
280
324
8.1
0.22
0.002
0.048
Dressing proportion, g/kg
527
534
505
542
10.6
0.39
0.042
0.16
Conformation score, EUROPb
4.88
5.00
4.38
4.99
0.257
0.25
0.15
0.32
Fat score, EUROPc
2.50
2.63
2.38
2.26
0.180
0.17
0.94
0.49
a
SEM = Standard error of mean. b Conformation: (1 = poorest, 15 = excellent). c Fat score: (1 = leanest, 5 =
fattest).
There were no significant interactions for final LW or LWG between feeding method and
concentrate level (Table 2). Feeding method had no effect on LWG but increasing
concentrate level led to a 6% improvement of daily LWG (P<0.05). Carcass gain was not
affected by feeding method but increasing concentrate allowance improved it 13%, on
average (P<0.01). There was also a significant interaction (P<0.05) between feeding method
and concentrate level for carcass gain. Carcass gain improved more with separate feeding
than with TMR-feeding when concentrate level was increased (Table 2).
Consistent with the findings of Petchey and Broadbent (1980), Caplis et al. (2005) and Keane
et al. (2006) there was no effect of mixing on LWG. On the contrary, Cooke et al. (2004)
reported that mixing of maize silage, grass silage, straw and concentrates resulted in a LWG
response of proportionately 0.15 for an intake increase of proportionately 0.04. A possible
explanation for the difference between the other findings and those of Cooke et al. (2004)
may be that the forage used by Cooke et al. (2004) included maize silage, straw and grass
silage whereas in other studies only grass silage was used. Regarding dairy cows, Yan et al.
(1998) have noted that when a benefit was obtained to TMR-feeding, forages other than grass
silage were offered.
106
Proceedings of the 5th Nordic Feed Science Conference
Ruminant nutrition
As in the present study, increasing concentrate allowance has improved growth in several
feeding experiments (e.g. Huuskonen et al. 2007; Manni et al. 2013). In the present study
observed increase in LWG was 13 and 39 g/d per kg increase in concentrate DMI for TMR
and separate feedings, respectively. This is roughly consistent with Huuskonen et al. (2007)
who reported an increase of 27 g/d in LWG per kg increase in concentrate DMI. There were
no significant interactions for feed or energy conversion between feeding method and
concentrate level (Table 2). However, due to differences in intake and growth parameters,
there were differences between feeding treatments. Both feed (kg DM/kg LWG) and energy
(MJ ME/kg LWG) conversion rates improved in separate feeding compared to TMR-feeding.
However, there were no significant differences between feeding methods when calculated per
carcass gain (Table 2). Furthermore, feed conversion (kg DM/kg carcass gain) improved and
energy conversion (MJ ME/kg LWG) declined with increasing concentrate proportion
(P<0.05).
The carcass weight of the bulls was 305 kg, on average, and close to pre-planned. Feeding
method had no effect on carcass weight but increasing concentrate allowance increased it by
9% (P<0.01). There was also a significant interaction (P<0.05) between feeding method and
concentrate level for carcass weight. There were no interactions for dressing proportion,
carcass conformation score or carcass fat score between feeding method and concentrate
level. Feeding method had no effect on dressing proportion, conformation score or fat score.
Concentrate level had no significant effects on conformation or fat score but increasing
concentrate allowance increased dressing proportion by 4% (P<0.05). As in the present
experiment Caplis et al (2005) and Keane et al. (2006) observed that feeding method
(separate vs. TMR) had no effects on slaughter traits of growing and finishing cattle. The
increasing effect of concentrate level on dressing proportion agrees with previous reports
(Caplis et al. 2005, Keane et al. 2006). In the present experiment, increasing concentrate
proportion did not improve the carcass conformation, consistent with Huuskonen et al. (2007)
and Randby et al. (2010) but contrary to Keane and Fallon (2001) and Caplis et al. (2005).
Conclusions
The present data showed that feeding a TMR increased feed intake of the bulls but had no
effect on growth or carcass traits. Increasing concentrate proportion increased growth
performance, carcass weight and dressing proportion of the bulls but had no effects on
carcass conformation score or carcass fat score.
References
Caplis, J., Keane, M.G., Moloney, A.P. & O'Mara, F.P., 2005. Effects of supplementary
concentrate level with grass silage, and separate or total mixed ration feeding, on
performance and carcass traits of finishing steers. Ir. J. Agric. Food Res., 44, 27–43.
Cooke, D.W.I., Monahan, F.J., Brophy, P.O & Boland, M.P., 2004. Comparison of
concentrates or concentrates plus forage in a total mixed ration or discrete ingredient format:
effects on beef production parameters and on beef composition, colour, texture and fatty acid
profile. Ir. J. Agric. Food Res., 43, 201–216.
EC 2006 Council Regulation (EC) No 1183/2006 of 24 July 2006 concerning the Community
scale for the classification of carcasses of adult bovine animals. The Official Journal of the
European Union L, 214, 1–6.
Proceedings of the 5th Nordic Feed Science Conference
107
Ruminant nutrition
Huuskonen, A., 2013. Performance of growing and finishing dairy bulls offered diets based
on whole-crop barley silage with or without protein supplementation relative to a grass
silage-based diet. Agric. Food Sci., 22, 424–434.
Huuskonen, A., Khalili, H. & Joki-Tokola, E., 2007. Effects of three different concentrate
proportions and rapeseed meal supplement to grass silage on animal performance of dairybreed bulls with TMR feeding. Livest. Sci., 110, 154–165.
Kaufmann, W., 1976. Influence of the composition of the ration and feeding frequency on pH
regulation in the rumen and on feed intake in ruminants. Livest. Prod. Sci., 3, 103–114.
Keane, M.G., 2010 A comparison of finishing strategies to fixed slaughter weights for
Holstein Friesian and Belgian Blue × Holstein Friesian steers. Ir. J. Agric. Food Res. 49, 41–
54.
Keane, M.G., Drennan, M.J. & Moloney, A.P., 2006. Comparison of supplementary
concentrate levels with grass silage, separate or total mixed ration feeding, and duration of
finishing in beef steers. Livest. Prod. Sci., 103, 169–180.
Keane, M.G. & Fallon, R.J., 2001. Effects of feeding level and duration on finishing
performance and slaughter traits of Holstein-Friesian young bulls. Ir. J. Agric. Food Res., 40,
145–160.
Manni, K., Rinne, M. & Huhtanen, P., 2013. Comparison of concentrate feeding strategies for
growing dairy bulls. Livest. Sci. 152, 21–30.
MTT 2014. Feed tables and nutrient requirements. MTT Agrifood Research Finland.
Retrieved 8 May 2014, from http://www.mtt.fi/feedtables.
Petchey, A.M. & Broadbent, P.J., 1980. The performance of fattening cattle offered barley
and grass silage in various proportions either as discrete feeds or as a complete diet. Anim.
Prod. 31, 251–257.
Phipps, R.H., Bines, J.A., Fulford, Rosemary J. & Weller, R.F., 1984. Complete diets for
dairy cows: a comparison between complete diets and separate ingredients. J. Agric. Sci.
Camb. 103, 171–180.
Randby, Å.T., Nørgaard, P. & Weisbjerg, M.R., 2010. Effect of increasing plant maturity in
timothy-dominated grass silage on the performance of growing/finishing Norwegian Red
bulls. Grass Forage Sci., 65, 273–286.
Yan, T., Patterson, D.C. & Gordon, F.J., 1998. The effect of two methods of feeding the
concentrate supplement to dairy cows of high genetic merit. Anim. Sci. 67, 395–403.
108
Proceedings of the 5th Nordic Feed Science Conference
Ruminant nutrition
Effect of salt addition to a barley concentrate on milking frequency, milk yield and feed
intake in automatic milking systems
M. Johansen, M. Larsen, P. Lund & M.R. Weisbjerg
Aarhus University, AU Foulum, Department of Animal Science, P.O. Box 50, 8830 Tjele,
Denmark
Correspondence: [email protected]
Introduction
In automatic milking systems (AMS) high amounts of concentrates are often offered in the
milking robot to achieve a high number of voluntary visits. But high intake of concentrates in
relative short time can affect the rumen environment negatively and increase the risk for
production diseases as rumen acidosis (Krause and Oetzel, 2006). Therefore, it is interesting
to investigate if some concentrates are so preferred that they increase the voluntary visits to
the milking robot, which would open for a potential reduction in concentrate offer. A direct
way to study this is by investigating cows’ response to different concentrates offered in the
milking robot on milking frequency, milk yield and feed intake. Most commercial compound
concentrates have a content of salt, but little is known about the influence on feed
preferences. Therefore, the objective of this study was to investigate how concentrates with
different contents of salt offered in the milking robot affect milking frequency, milk yield and
feed intake.
Materials and methods
The experiment was conducted at the Danish Cattle Research Centre (KFC), Foulum,
Denmark, in a loose housing system with a DeLaval milking robot, to which the cows had
free access. Forty-eight Jersey cows, 16 primiparous and 32 multiparous, were used for the
experiment. The cows were 173 ± 128 (mean ± SD) days from calving at the beginning of the
experiment. The cows were divided into 12 blocks according to parity and days from calving
and randomly assigned to one of the four treatments in a 4x4 Latin square. Each period
continued for one week without any adaptation, and in total the experiment lasted for four
weeks. The four treatments were a standard pelletized robot concentrate with a content of 0.7
percent salt used as a control (CON) and pelletized barley with 0 (B0), 0.7 (B0.7) or 1.4
(B1.4) percent salt, respectively. Max concentrate offer was 3 kg daily, divided over daily
robot visits according to time interval from last milking. Concentrate allowance was fed out
at a rate of 0.4 kg/minute. At the feeding lane the cows had free access to a mixed ration
(MR) composed of 31.5% grass silage, 25.5% maize silage, 11.2% rapeseed cake, 11.2%
sugar beet pulp, 9.9% barley, 9.9% NaOH treated wheat, 0.7% minerals and vitamins and
0.1% salt on dry matter basis. Milking frequency, amounts of concentrates offered in the
milking robot, orts of concentrates left, milk yield and feed intake of the MR were registered
daily at cow level. Milk composition at each milking during 48 h was analyzed each week.
All statistical analyses were performed using repeated measures within the MIXED
procedure of SAS (SAS® version 9.3; SAS, 2010). The model included cow nested in parity,
treatment, parity (primi- or multiparous), day within period (day), and the two-way
interactions between treatment and parity and between treatment and day. Measures within
cow x treatment were handled as repeated measures with a compound symmetry (cs)
covariance structure. The denominator degrees of freedom were estimated by the
“KenwardRoger” method. Significance was determined by P ≤ 0.05.
Proceedings of the 5th Nordic Feed Science Conference
109
Ruminant nutrition
Results
One multiparous cow was removed from the data due to laminitis during the experiment.
Weight of the cows was unaffected by the experiment, since the mean (±SD) weight at the
beginning was 492.6 (±39.1) kg and at the end was 492.1 (±40.6) kg.
The milking frequency was unaffected by treatment, but multiparous cows had a higher
milking frequency than primiparous cows (Table 1). The amount of concentrate offered in the
robot was lower for the three barley treatments compared with the control treatment. There
was an interaction between treatment and parity (P = 0.01) on orts of concentrates, as the
primiparous cows had similar orts for B1.4 and the control treatment, whereas multiparous
cows had 3.5 times higher orts for the B1.4 treatment compared with the control.
Table 1 Effect of treatments on milking frequency, amount of concentrate offered in the robot, amount of
concentrate left in the robot, milk yield and intake of the mixed ration at the feeding lane
Treatmenta
CON
B0
B0.7
P-valueb
B1.4
SEMc
Milking frequency (times/day)
Primiparous
2.39
2.41
2.42
2.49
0.07
Multiparous
3.10
3.15
3.08
3.03
0.05
Amount offered (kg/day)
Primiparous
2.83
2.71
2.61
2.66
0.05
Multiparous
2.93
2.73
2.58
2.73
0.04
Amount orts (kg/day)
Primiparous
0.09
0.18
0.13
0.09
0.05
Multiparous
0.10
0.19
0.17
0.35
0.04
Milk yield (L/day)
Primiparous
19.3
17.9
18.2
18.4
0.27
Multiparous
23.5
22.8
22.9
22.5
0.19
ECM (L/day)
Primiparous
25.7
24.3
24.3
24.7
0.36
Multiparous
31.4
30.8
30.9
30.2
0.26
Intake of MR (kg DM/day)
Primiparous
13.8
13.3
13.4
13.3
0.19
Multiparous
17.7
17.3
17.5
17.4
0.14
Treat
Par
Treat x Par
Day
Day x Treat
0.97
<0.001
0.37
0.16
0.11
<0.001
0.20
0.56
0.27
0.006
0.05
0.01
0.01
0.36
0.48
<0.001
<0.001
0.22
0.03
<0.001
0.002
<0.001
0.18
0.03
<0.001
0.12
<0.001
0.89
0.33
0.50
a
CON: standard robot concentrate; B0: pelletized barley; B0.7: pelletized barley with 0.7% salt; B1.4: pelletized
barley with 1.4% salt; b Treat.: treatments; Par.: parity; Day: day within period; c Standard error of the mean
The milk yield was influenced by treatment and parity (Table 1). Multiparous cows had a
higher milk yield than primiparous cows and for both parities the milk yield was lower at the
three barley treatments compared with the control treatment (P < 0.001). For energy corrected
milk (ECM), based on the weekly milk analyses, the differences between the treatment means
were as found for milk yield. Both for milk yield and ECM there was an interaction between
110
Proceedings of the 5th Nordic Feed Science Conference
Ruminant nutrition
day within period and treatment (Table 1). At day one in the treatment period there was no
difference in milk yield between treatments, but at day two and the following days, the milk
yield was lower for the three barley treatments compared with the control treatment.
The feed intake of the MR at the feeding lane was unaffected by treatments (Table 1), but the
multiparous cows had a higher feed intake than the primiparous cows.
Discussion
The content of Na in the MR was 3.1 g/kg DM (40% above the norm) to guarantee the
requirement of Na to be fulfilled, by which possible effects would be due to taste and feed
preference instead of unfulfilled requirements.
The milking frequency was unaffected by the treatments, which indicated that the cows’
preferences for barley does not differ from their preference for the control concentrate, and
moreover are not affected by salt addition. Barley is generally believed to be a feedstuff cows
like to eat, and Klopfer et al. (1981) found barley to be preferred highest in a choice test
comparing 20 different feedstuffs. In a choice test comparing pelletized barley with the same
control concentrate as used in this experiment, the control concentrate was preferred highest
(Primdal et al., 2014). This indicates that cows have high preferences for both barley and the
control concentrate.
Addition of salt did not affect milking frequency, but the amount of orts left in the robot
indicate that multiparous cows had lower preference for the pelletized barley with the high
content of salt since they left more orts compared with the other treatments. In the choice test
by Klopfer et al. (1981) salt blocks were preferred lowest, which indicate that cows do not
prefer the taste of pure salt. Nombekela et al. (1994) found by testing preferences for the
primary tastes sweet, salty, sour and bitter in total mixed rations, that the salty diet was
ranked very low as the probability of the salty diet to be chosen first when all diets were
presented was only 1%. Further, Nombekela et al. (1994) found that the daily dry matter
intake was unaffected even though the most preferred diets one by one were removed. This
shows that feedstuffs that have ranked low in a preference trial not necessarily influence the
feed intake or the milking frequency, when the cows do not have a choice.
The amount of concentrate offered in the robot differed between treatments even though the
milking frequency was unaffected. The average difference between the amounts offered was
0.2 kg/day, and Halachmi et al. (2005) found no effect at the milking frequency by reducing
the amount of concentrates offered in the milking robot by 1.2-1.5 kg/day. The lower offer of
concentrates could be a result of briefer visits in the robot on the test treatments compared
with the control treatment, although not studied. The lower milk yield on the test concentrates
compared with the control concentrate could partly be due to a lower intake of energy
because of a lower offer of the test concentrates in the milking robot.
Conclusion
Milking frequency was not affected by the offer of different concentrates in the milking robot
and no effects were found by salt addition to barley.
Proceedings of the 5th Nordic Feed Science Conference
111
Ruminant nutrition
References
Halachmi, I., Ofir, S. & Miron, J., 2005. Comparing two concentrate allowances in an
automatic milking system. Journal of Animal Science 80, 339-343.
Klopfer, F.D., Kilgour, R. & Matthews, L.R., 1981. Paired comparison analysis of
palatabilities of twenty foods to dairy cows. Proceedings of the New Zealand Society of
Animal Production 41, 242-247.
Krause, K.M. & Oetzel, G.R., 2006. Understanding and preventing subacute ruminal acidosis
in dairy herds: A review. Animal Feed Science and Technology 126, 215-236.
Nombekela, S.W., Murphy, M.R., Gonyou, H.W. & Marden, J.I., 1994. Dietary preferences
in early lactation cows as affected by primary tastes and some common feed flavors. Journal
of Animal Science 77, 2393-2399.
Primdal, L., Johansen, M. & Weisbjerg, M.R., 2014. Do dairy cows have preferences for
different concentrate feeds? Animal Production Science [submitted, in process].
SAS, 2010. Statistical Analysis System. User’s guide. SAS Institute Inc., Cary, NC.
112
Proceedings of the 5th Nordic Feed Science Conference
Ruminant nutrition
The evaluation of efficiency of different mineral additives on milk yield and quality in
dairy cows
V. Jokubauskienė, V. Špakauskas
Lithuanian University of Health Sciences, Tilzes str. 18, 47181 Kaunas, Lithuania
Correspondence: [email protected]
Introduction
Scientific and technological progress in agriculture is of great importance to the performance
of livestock. Higher milk yields, better quality of milk, higher gains and increased fertility
rates are related with nutrition. A well balanced diet is one of the main factors affecting the
health of livestock. A lot of trace elements supplements of various chemical compositions are
created to improve performance of livestock. Various researchers (Griffiths et al., 2007;
Karkoodi et al., 2012; Krys et al., 2009; Weiss et al., 2010) have tried to find out how trace
minerals supplements impact on animal performance. The results have been diverse. This
indicates that all factors, which are affecting the nutritional needs of organisms, are not fully
understood. Therefore, the aim of our study was to evaluate the efficiency of different trace
mineral additives on cows’ milk yield and milk quality.
Materials and Methods
The experiment was carried out with cows in the dairy farm of Kaišiadoriai. Sixty clinically
healthy cows were selected for the investigation taking into consideration time of calving,
nutrition and state of health. The experimental cows were divided into three groups, of 20
cows each. Cows in the first groups were fed a ration with inorganic mineral additives (IMA),
which was made from calcium, magnesium and potassium chlorides, sodium dihydrogen
phosphate, cobalt carbonate, copper, zinc and manganese sulphates, potassium iodide,
sodium selenite. The cows of the second group received a complexed mineral additive
(CMA), which was made from inorganic (magnesium and potassium chlorides, sodium
dihydrogencarbonate, manganese sulphate, potassium iodide and sodium selenite) and
organic (zinc, copper and cobalt gluconates, and potassium lactate) salts. The last group –
constituted the Control cows, which did not receive any mineral additives, only licks of
sodium chloride. Tested minerals supplements were fed to the IMA and CMA cows every
other day for 3 months at levels indicated in Table 1.
Samples of milk were taken once every month, during control milking. Amount of milk yield
(kg/d), fats, proteins, lactose and somatic cell counts (SCC) in milk were determined once
every month. Milk yield was measured on farm, and the indicators of composition and SCC
were determined in the laboratory of ‘Pieno tyrimai’, by use of a Lacto Scope FTIR
(FT1.0.2001) and Soma Scope (CA-3A4, 2004), respectively. Subclinical mastitis was
detected by using the California Mastitis Test (CMT).
Results and Discussion
Treatment groups IMA and CMA produced more milk (4.84% and 6%, respectively)
compared to the control cows. Cows, which received CMA, also produced more milk
(1.25%), compared with cows in the IMA group (Fig. 1.).Milk fat contents of IMA, CMA
and control cows ranged from 3.72 – 4.3%, 3.55 – 4.1% and 3.79 – 4.11%. Milk fat of IMA
and CMA cows tended to
Proceedings of the 5th Nordic Feed Science Conference
113
Ruminant nutrition
Table 1 The amount (g) of mineral additives fed to dairy cows once every day
The composition of IMA
The composition of CMA
Calcium chloride
30.0 g
Calcium lactate, 19 proc.
28.00g
Magnesium chloride
5.00 g
Magnesium chloride
5.00g
Sodium dihydrogen phosphate
4.00 g
Sodium dihydrogen phosphate
4.00 g
Potassium chloride
0.50 g
Potassium chloride
0.50 g
Cobalt carbonate
0.03g
Cobalt gluconate, 12–13.5 proc.
0.03 g
Copper sulphate
0.48 g
Copper gluconate, 13 – 15 proc.
0.85 g
Zinc sulfate
1.66 g
Zinc gluconate, 12 – 14 proc.
3.86 g
Manganese sulphate
1.66 g
Manganese sulphate
1.66 g
Potassium iodide, stabilized
0.02 g
Potassium iodide, stabilized
0.02 g
Sodium selenite
0.01 g
Sodium selenite
0.01 g
23,5
IMA
CMA
Control
23,0
22,5
22,0
kg
21,5
21,0
20,5
20,0
19,5
19,0
18,5
0-Aug
1-Sept
2-Oct
3-Nov
Months
Figure1 The dynamic of milk yield of cows before and after supplementation. The milk yield of cows tended to
decrease after 3 months following treatment. The cows in IMA and CMA group produced more milk compared
to the control cows.
increase after two months following the feeding of mineral additives. Milk fat content of
cows, which received IMA, was higher (4.09% ± 0.135; P < 0.05) compared with control
cows (3.94 ± 0.072) and was 3.64% higher relative to cows, which received CMA (P<0.05)
(Fig. 2).The amount of milk protein of treatment and control cows did not differ (P>0,05), but
cows supplemented with CMA, produced more milk proteins (0,61%), compared with those
receiving IMA (Fig. 3.). Milk lactose content of IMA, CMA and control cows ranged from
4.5 to 4.63, 4.3 – 4.68 and 4.43 – 4.60%, respectively. Milk lactose content tended to increase
after 1, 2, 3 months following administration of CMA, and after 3 months, it was 1,08%
higher compared with cows in the IMA group, P > 0,05.
114
Proceedings of the 5th Nordic Feed Science Conference
Ruminant nutrition
4,20
IMA
CMA
Control
4,15
4,10
%
4,05
4,00
3,95
3,90
3,85
3,80
0-Aug
1-Sept
2-Oct
3-Nov
Months
Figure 2 The dynamic of milk fats of cows before and after supplementation. Milk fat content of cows in IMA
group was higher compared to the control and CMA cows.
3,42
IMA
CMA
Control
3,40
3,38
%
3,36
3,34
3,32
3,30
3,28
3,26
3,24
0-Aug
1-Sept
2-Oct
3-Nov
Months
Figure 3 The dynamic of milk protein concentration of cows before and after supplementation. The amount of
milk protein between treatment and control cows did not differ.
The least SCC in the milk of treatment cows was determined after 3 months, following
administration of the mineral supplements. After 1 and 2 months, following administration of
CMA, the SSC tended to decrease and was 8.7% and 10.7% lower than for the control cows
and 13.8% and 23.29% lower than for cows on the IMA supplement.
Proceedings of the 5th Nordic Feed Science Conference
115
Ruminant nutrition
4,66
IMA
CMA
Control
4,64
4,62
%
4,60
4,58
4,56
4,54
4,52
4,50
4,48
0-Aug
1-Sept
2-Oct
3-Nov
Months
Figure 4 The dynamic of milk lactose of cows before and after supplementation. Milk lactose of cows of CMA
and IMA groups tended to increase after 3 months following supplementation, compared with control cows.
320
IMA
CMA
Control
1000's/ml
300
280
260
240
220
200
0-Aug
1-Sept
2-Oct
3-Nov
Months
Figure 5 The dynamic of SCC in milk of cows before and after supplementation. The SCC in milk of treatment
cows tended to decrease after 3 months following supplementation. Cows of CMA group had lower amount of
SCC in milk compared to cows in IMA group.
Conclusions
Mineral additives have a positive effect on milk yield and milk quality in dairy cows.
Complexed mineral additives seem to be more effective than inorganic mineral additives on
productivity and milk quality of dairy cows.
References
Griffiths, L. M., Loeffler, S. H., Socha, M. T., Tomlinson, D. J. & Johnson, A. B., 2007.
Effects of supplementing complexed zinc, manganese, copper and cobalt on lactation and
reproductive performance of intensively grazed lactating dairy cattle on the South Island of
New Zealand. Anim. Feed Sci. Technol., 137, 69–83.
116
Proceedings of the 5th Nordic Feed Science Conference
Ruminant nutrition
Karkoodi, K. & Chamani, M., 2012. Effect of Organic Zinc, Manganese, Copper, and
Selenium Chelates on Colostrum Production and Reproductive and Lameness Indices in
Adequately Supplemented Holstein Cows. Biol. Trace Elem. Res., 146, 42 – 46.
Krys, S., Lokajová, E., Podhorský, A. & Pavlata, L., 2009. Microelement Supplementation in
Dairy Cows by Mineral Lick. Acta Vet. Brno., 78, 29–36.
Weiss, W. P., Pinos-Rodríguez, J. M. & Socha, M. T., 2010. Effects of feeding supplemental
organic iron to late gestation and early lactation dairy cows. J. Dairy. Sci., 93, 2153-2160.
Proceedings of the 5th Nordic Feed Science Conference
117
Ruminant nutrition
Selenium supplementation by addition of sodium selenate with silage additive
A. Seppälä1, Y. M. Albarran2, H. Miettinen3, M. P. Siguero2, E. Juutinen1,4 & M. Rinne1
1
MTT Agrifood Research Finland, Animal Production Research, 31600 Jokioinen, Finland;
2
Department of Analytical Chemistry, Faculty of Chemistry, Complutense University of
Madrid, Madrid Ciudad Universitaria s/n 28040 Madrid, Spain; 3Kemira Oyj, P.O. Box 330,
00101 Helsinki, Finland, current: Nöykkiölaaksontie 27C6, 02330, Espoo 4Current address:
Atria Ltd, Finland
Correspondence: [email protected]
Introduction
Despite three decades of selenium enriched fertilizers in Finland, the problem with low
selenium-bioavailability persists. Selenium content in agricultural soils in Finland has not
changed significantly during those years and only 4% of the total selenium in the soil is
available to plants (Eurola et al. 2011). Thus efforts should continue to prevent possible
harmful consequences of low selenium bioavailability.
Although the most severe symptoms of selenium deficiency in cattle (white muscle disease)
are quite rare nowadays (Eurola et al. 2011), the less-precise symptoms such as muscular
weakness of the newborn, reduced weight gain, diarrhoea or decreased fertility (Koller and
Exon 1986) may be linked to selenium deficiency. Further sufficient selenium is needed to
ensure optimal udder health (Jukola et al. 1996). Smith et al. (1985, 1997) found over 30%
reduction in the number of cases of clinical mastitis and nearly a 70% reduction in the
number of high somatic-cell-count cows as a result of supplementation with Se and vitamin
E. Flocks of dairy sheep having high incidence rates of clinical mastitis (> 10%) were found
to have lower (P<<0.001) blood selenium and serum vitamin A concentrations than flocks
with low rates of incidence of mastitis (<3%) (Giadinis et al. 2011). These authors suggested
that the selenium status of ewes may possibly be used to indicate animals at risk of
developing clinical mastitis (Giadinis et al. 2011). The low selenium content of cows' milk
(less than 0.15 mg kg-1 DM) produced on organic farms compared to non-organic milk
(average 0.23 mg kg-1 DM) reflects the low level of selenium in diets on those farms (Eurola
et al. 2011). High cost of each case of clinical mastitis (270-670 € per case for dairy cows;
Heikkilä et al. 2010) encourage to ensure sufficient selenium for lactating mammals.
Although selenium-enriched fertilizers exist they are not always used, as the benefit of
selenium is coming indirectly via animal production. Furthermore, in organic farming, the
use of compound fertilizers is not allowed. The selenium content of organic silages, or silages
produced without compound fertilizers, was only one-tenth that found in conventionally
produced silages from grassland fertilized with selenium: 0.02 vs. 0.2 mg Se kg-1 DM (Eurola
et al. 2011). The selenium content of such silages is not sufficient to fulfill the selenium
requirement of cattle (0.1 mg Se kg-1 DM according to MTT (2013) or 0.3 mg Se kg-1 DM for
dairy cattle according to NRC (2001)). Organically produced cereal grains are also low in
selenium, so additional selenium sources are needed to ensure sufficient selenium supply to
livestock and humans.
Maintenance feeding of ruminants and horses is typically based on forage, typically
supplemented with some minerals. Suckler cows and ewes may be fed forage as a sole energy
source for more than half of the indoor-feeding period, typically until the last weeks of
pregnancy when concentrates are normally added to the diet. The dosage of mineral feeds
may be challenging if silage is fed ad libitum and minerals are offered separately without
118
Proceedings of the 5th Nordic Feed Science Conference
Ruminant nutrition
individual control. It is difficult to ensure a sufficient, but not toxic, level of selenium in those
situations if the forage itself is low in selenium. Supplementation of individual animals may
also not be a practical solution in large herds, and therefore forage with sufficient selenium
content would be needed, especially for these animal groups.
Silage additives are evenly sprayed on to the grass material prior to ensiling, and adding
selenium to the silage additive might be a practical solution to ensure that the required dose
of selenium is applied safely and evenly in the silage. An ensiling experiment was conducted
to explore the effect of a sodium selenate-containing silage additive on the selenium content
of silage and on the content of different forms of selenium.
Materials and Methods
The study was performed at MTT Agrifood Research Finland in Jokioinen in 2011. A
timothy (Phleum pratense) -meadow fescue (Festuca pratensis) sward was fertilized using
compound fertilizers without added selenate during summer 2011. The second cut was
harvested on 2 August 2011 and prewilted for three hours before harvesting with a precision
chopper.
Five batches of grass material were weighed and each batch was given the additive treatment
according to the design of the experiment (Table 1). Each batch of grass was mixed manually
during and after applying the additive. Three replicate silos (12 L, 7 kg grass per silo) were
filled for each additive. The silos were opened and sampled after a 107-day ensiling period.
Table 1 Experimental treatments with additive dose level of 6 g kg-1
Abbreviation
Added selenium to grass, mg kg-1
Composition
Control
Water
W50
Water with 50 mg kg-1 sodium selenate
A10
-1
AIV Ässä* with 10 mg kg sodium selenate
-1
Sodium selenate
Selenium
-
-
0.30
0.125
0.06
0.025
A50
AIV Ässä with 50 mg kg sodium selenate
0.30
0.125
A500
AIV Ässä with 500 mg kg-1 sodium selenate
3.00
1.254
*Composition of AIV Ässä: 590 g formic acid, 200 g propionic acid, 45 g ammonium formate, 25 g benzoic
acid/sorbate and 140 g water per kg.
The DM concentration was determined by drying at 105 °C for 16 h. Ash content was
determined using a standard method of AOAC (1990; method 942.05). Nitrogen (N) content
was determined by the Dumas method (AOAC method 968.06) using a Leco FP 428 nitrogen
analyzer. Crude protein content was calculated as 6.25 × N content. Concentration of neutral
detergent insoluble fibre (NDF) was determined according to Van Soest et al. (1991) using
sodium-sulphite, without amylase and presented ash free.
Silages were analysed for volatile fatty acids (VFA) according to Huhtanen et al. (1998),
lactic acid according to Haacker et al. (1983), water soluble carbohydrates (WSC) according
to Somogyi (1945) and ammonia according to McCullough (1967). Oven DM concentration
of silage was corrected for the loss of volatiles according to Huida et al. (1986).
The determination of the total selenium content was carried out by inductively coupled
plasma mass spectrometry (ICP-MS) (HP 7700 Agilent Technologies, Santa Clara,
California, United States) after a microwave acid digestion. Each dry sample (50 mg) was
Proceedings of the 5th Nordic Feed Science Conference
119
Ruminant nutrition
weighed and 1 mL of nitric acid (650 g kg-1) and 0.3 mL of hydrogen peroxide (350 g kg-1)
were added. The samples were subjected to acid digestion in a microwave oven (0-130°C for
15 min followed by 130°C for 10min). After digestion, the resulting extracts were diluted
with Milli-Q water up to a final volume of 12 mL. Samples were analysed in triplicate.
Identification of selenium species of silage samples was carried out by high-performance
liquid chromatography coupled to inductively coupled plasma mass spectrometry (HPLCICP-MS) (HP 7700 Agilent, Santa Clara, California, United States) after an enzymatic
hydrolysis with protease. The recovery of selenium present in the samples after the enzymatic
hydrolysis was more than 95%. The extraction of selenium species from 50 mg of dry sample
was performed by 2 min sonication after the addition of 3 mL of buffer Tris-HCl (30mM, pH
7.5) and 0.020 mg of Protease type XIV isolated from Streptomyces griseus, ≥3.5 units/1mg
solid (purchased from Sigma-Aldrich, St. Louis, USA). The extracts were centrifuged at
11000 rpm for 15 min and then filtered by means of using 0.20 µm nylon filters. The
supernatants were analysed by anion exchange chromatography coupled to ICP-MS. The
concentrations of selenium species were determined by monitoring 77Se, 78Se, and 82Se
isotopes by comparison to known standards solutions. The identification was performed by
comparing retention times with those of standard solutions. The standard included
selenocysteine (SeCys2), seleno-methyl-selenocysteine (SeMetSeCys), selenite Se(IV),
selenomethionine (SeMet) and selenate (Se(VI)).
Statistical analyses were performed using SAS GLM-procedure. Treatment effects on
fermentation quality and selenium content of the silages were tested by analysis of variance
and differences between treatments (lsmeans) were compared using the Tukey test.
Results and Discussion
The harvested grass prior to ensiling had a dry matter content of 283 g kg-1. In the dry matter,
the contents of ash, crude protein, WSC and NDF were 113, 175, 147 and 536 g kg-1,
respectively.
The formic acid-based silage additive ‘AIV Ässä’ restricted fermentation compared to the
Control and W0 treatments, as AIV Ässä silages had more WSC, less fermentation products
and less ammonium N than the silages without ‘AIV Ässä’ (Table 2). Selenium dosed within
the additive was evenly sprayed on to the silage with the AIV Ässä product, as indicated by
the total selenium content of the silages being close to the theoretical concentration (Table 3).
The largest deviation (45%) was observed in the W50 treatment. Nearly all (95%) of the total
selenium was detected as sodium selenate and no other forms were detected.
The Control silage had a total selenium content of 0.069 mg kg-1 DM (Table 3), which is
clearly below the selenium requirement of cattle and sheep (0.1 mg Se kg-1 DM according to
MTT (2013) or 0.3 mg Se kg-1 DM according to NRC for dairy cattle (2001)). The lowest
level of addition of sodium selenate produced silage that had a content of selenium (0.159 mg
kg-1) that was sufficient to prevent deficiency symptoms in cattle. The intermediate level of
addition resulted in silage that had selenium content 65% above the NRC (2001)
recommendation for dairy cattle, but this was still lower than the highest permitted selenium
content in feed. The selenium content should not exceed 0.5 mg kg-1 in a feed ration having
12% DM content (MMM 43/2005), which is equal to 0.568 mg kg-1 DM. The highest
application level resulted in selenium content in silage that was 7.4-times higher than the
maximum allowed level.
120
Proceedings of the 5th Nordic Feed Science Conference
Ruminant nutrition
Table 2 Fermentation quality of the silages (units g kg-1 DM unless otherwise stated)
Control
Dry matter, g kg-1
W50
288
pH
A10
288
4.20
b
A50
279
4.20
b
A500
289
4.30
a
SEM
288
4.27
ab
0.8
4.32
a
Treatment
effect
ns
0.018 <0.001
Lactic acid
99.7a
97.8a
Acetic acid
21.18a
21.07a
9.65b
9.51b
9.29b
0.458 <0.001
c
c
b
b
a
0.211 <0.001
61.2b
66.5b
41.2c
Propionic acid
0.21
VFA4-VFA61
0.28
0.33
0.16
0.13
0.4
14.0c
14.1c
62.4b
60.5b
80.1a
a
a
b
b
b
2
WSC
-1
Ammonium N g kg N
65.2
0.15
69.0
2.94
20.8
2.89
16.4
2.29
4.14
28.0
<0.001
0.087 ns
1
2.38
<0.001
3.64
<0.001
2
VFA4-VFA6 = the sum of butyric, isobutyric, valeric, isovaleric and caproic acids; Water soluble
carbohydrates; SEM=Standard error of the mean Treatment means within the same row but without the same
superscript are statistically different (p<<0.05, Tukey test).
Table 3. Expected and analysed selenium content in the silages. Selenium content expressed as mg Se kg-1 DM.
Treatments are explained in Table 1
SEM4
Treatment
effect
4.199 c
0.125
<0.001
4.457 b
0.193
<0.001
Control
W50
A10
A50
A500
Expected selenium content1
-
0.507
0.157
0.507
4.453
Analysed total selenium2
0.069a
0.920 b
0.159 a
0.496 ab
Selenium detected as
selenate3
0.000 a
0.667 a
0.156 a
0.486 a
Treatment means within same row without same superscript are statistically different (p<<0.05, Tukey test);
1
Calculated as the sum of the selenium content of the Control silage and the amount of selenium theoretically
provided with the additive treatment; 2Determined by inductively coupled plasma mass spectrometry (ICP-MS);
3
Determined by high-performance liquid chromatography coupled to inductively coupled plasma mass
spectrometry; 4SEM=Standard error of the mean.
There are several aspects to be taken into account when optimizing the level of selenium
inclusion in the silage additive. The inclusion should be sufficient for the selenium to have
positive effects on animal health, but there is a need to ensure the amount is still safe, and
takes account particularly of situations where the grass has also received selenium-enriched
fertilizer. The optimization is complicated by the fact that the dosage recommendations of
silage additives are typically expressed on fresh-matter basis and the DM content of the grass
typically varies from 200 to 400 g kg-1. Two possible scenarios were calculated assuming
selenium content in grass either 0.03 or 0.37 mg kg-1 DM, DM in grass either 400 or 200 g
kg-1 and additive dosage level 5 l/t. The lower level of inclusion (4.67 mg selenium kg-1
additive, 11 mg kg-1 sodium selenate) is sufficient to prevent selenium deficiency in cattle
having silage as their sole feed, despite low inherent selenium content in the grass. The
maximum allowed inclusion of selenium (6.61 mg selenium kg-1 additive, 15 mg kg-1 sodium
selenate) was calculated by assuming low dry matter content of the grass and high inherent
Proceedings of the 5th Nordic Feed Science Conference
121
Ruminant nutrition
selenium content. These two cases demonstrate that the selenium content of the silage
additive can vary only within a relatively narrow range.
Conclusions
Addition of sodium selenate to silage additive provided a controlled way to add selenium to
the diet of forage-fed animals. Silage additives were applied evenly to the material at ensilage
so as to ensure successful ensiling; this would, at the same time, allow the possibility of
distributing selenium evenly to animals consuming the silage without additional labour costs.
The volume of added selenium was so small that it could be included in the acid-based
additive without changing the dosage recommendations. The incremental addition of
selenium into the additive could take place in the controlled environment of the factory, thus
posing no danger to farmers or consumers. The level of selenium inclusion can be chosen so
that selenium deficiency symptoms are prevented but without the risk of exceeding the
highest allowed selenium content of the diet, taking account of grass material that has also
received selenium enriched fertilizers.
References
AOAC, 1990. Official Methods of Analysis. Association of Official Analytical Chemists,
Inc., Arlington, VA. 1298 p.
Eurola, M., Alfthan, G., Ekholm, P., Root, T., Suoniitty, T., Venäläinen, E.-R. & Ylivainio,
K., 2011. Seleenityöryhmän raportti 2011. MTT Raportti 35, 34 s.
Giadinis, N.D., Panousis, N., Petridou, E.J., Siarkou, V.I., Lafi, S.Q., Pourliotis, K.,
Hatzopoulou, E. & Fthenakis, G.C., 2011. Selenium, vitamin E and vitamin A blood
concentrations in dairy sheep flocks with increased or low clinical mastitis incidence. Small
Rumin. Res. 95, 193-196.
Haacker, K., Block, H.J. & Weissbach, F. 1983. Zur kolorimetrischen
Milchsäurebestimmung in Silagen mit p-Hydroxydiphenyl. [On the colorimetric
determination of lactic acid in silages with p-hydroxydiphenyl]. Arch.Tierernähr. 33, 505512.
Heikkilä, A-M., Nousiainen, J. & Pyörälä, S. 2010. Kallis utaretulehdus. In: Toim. Anneli
Hopponen. Maataloustieteen Päivät 2010, Suomen maataloustieteellisen seuran tiedote 26, 4
p. Available at: http://www.smts.fi/jul2010/esite2010/089.pdf (cited 18 December 2013).
Huhtanen, P.J., Blauwiekel, R. & Saastamoinen, I. 1998. Effects of intraruminal infusions of
propionate and butyrate with two different protein supplements on milk production and blood
metabolites in dairy cows receiving grass silage based diet. J. Sci. Food Agric. 77, 213-222.
Huida, L., Väätäinen, H. & Lampila, M. 1986. Comparison of dry matter contents in grass
silages as determined by oven drying and gas chromatographic water analysis. Ann. Agric.
Fenn. 25, 215-230.
Jukola, E., Hakkarainen, J., Saloniemi, H. & Sankari, S. 1996. Blood Selenium, Vitamin E,
Vitamin A, and β-Carotene Concentrations and Udder Health, Fertility Treatments, and
Fertility. J. Dairy Sci. 79, 838–845.
McCullough, H. 1967. The determination of ammonia in whole blood by direct colorimetric
method. Clinica Chimica Acta 17, 297-304.
122
Proceedings of the 5th Nordic Feed Science Conference
Ruminant nutrition
Koller, L.D. & Exon, J.H. 1986. The two faces of selenium – deficiency and toxicity – are
similar in animals and man. Canad. J. Vet. Res. 50, 297-306.
MMM 43/2005. Maa- ja metsätalousministeriön asetus rehun lisäaineista ASETUS nro 43/05
Liite 1. Valmistajakohtaisesti hyväksyttävät sekä jakelu- ja käyttörajoituksien alaiset
lisäaineet.
MTT 2013. Feed tables and nutrient requirements. Available at:
https://portal.mtt.fi/portal/page/portal/Rehutaulukot/feed_tables_english (Cited 13 November
2013).
NRC 2001. Nutrient Requirements of Dairy Cattle. Seventh Revised Edition, National
Research Counsil, Washington D.C.
Somogyi, M. 1945. A new reagent for the determination of sugars. J. Biol. Chem. 160, 61-68.
Smith, K.L., Hogan, J.S. & Weiss, W.P. 1997. Dietary vitamin E and selenium affect mastitis
and milk quality. J. Anim. Sci. 75, 1659-1665.
Smith, K.L., Conrad, H.R., Amiet, B.A., & Todhunter, D.A. 1985. Incidence of
environmental mastitis as influenced by dietary vitamin-e and selenium. Kieler
Milchwirtschaftliche Forschungsberichte 37,482-486 (ref. Smith et al. 1997).
Van Soest, P.J., Robertson, J.B. & Lewis, B.A. 1991. Methods for dietary fibre, neutral
detergent fibre and nonstarch polysaccharides in relation to animal nutrition. J. Dairy Sci. 74,
3583-3597.
Proceedings of the 5th Nordic Feed Science Conference
123
Ruminant nutrition
Factors influencing the production value of forage protein
P. Huhtanen
Swedish University of Agricultural Sciences (SLU), Department of Agricultural Research for
Northern Sweden, S-901 83 Umeå, Sweden.
Introduction
Protein is the most expensive main nutrient in the diets of dairy cows. Increased prices of
protein feeds, environmental concerns of increased N emissions with increased protein
feeding and ethical concerns of using high quality proteins that could directly be used as
human food or utilized by monogastric animals have increased interest of improving
utilization of forage protein for ruminants. Development of new feed protein evaluation
systems in 1980’s that are based the supply of absorbed amino acids from microbial protein
and feed protein form a more comprehensive basis for evaluation of feed protein value.
Increasing the supply of amino acids from forages, especially of undegraded protein (RUP),
has been a subject of intensive research. Considerable differences in the predicted supply of
RUP from forages have been reported from studies based mainly on the in situ (nylon bag)
technique or chemical analysis of N fractions. However, these observations have very seldom
been validated in milk production studies or digesta flow studies in cannulated animals. The
objective of this paper is to discuss different strategies in increasing the supply of absorbed
amino acids (AAT) from forages.
Nitrogen fertilization
Increasing N fertilization of leys was a widely used strategy in 1970’s to increase protein
supply for dairy cows. However, this strategy was not challenged by comparing milk
production responses to increased crude protein (CP) supply from increased N fertilization
and high quality protein supplements. In a later study (Shingfield et al., 2001), increasing
fertilization from 50 to 100 kg/ha increase silage CP concentration from 120 to 150 g/kg DM.
Increased CP from silage had no influence on milk or protein yield when no protein
supplements were fed despite very low dietary CP and milk urea concentrations (121 g/kg
DM, 1.4 mM). However, when the low CP silage was supplemented with rapeseed expeller to
increased dietary CP to the same level as the high CP silage without supplementary protein
milk protein yield increased 120 g/d. Interestingly the response to rapeseed expeller was the
same (115 g/d) with high CP silage. The results of this study indicate that the cows were
responsive to increased protein supply, but the supply of utilizable protein could not be
increased by N fertilization. Both silages were well-preserved and had similar DM
concentration (340 g/kg). Consistently with production study increasing N fertilization of
grass ley (40, 80 and 120 kg/ha) had no influence on duodenal NAN flow study in duodenally
cannulated animals (Vanhatalo and Toivonen, 1993). It can be concluded that grass
fertilization should be optimized according to yield with no extra benefits from increased N
fertilization.
Grass maturity
It is well-known that harvesting grass silage at an earlier stage of maturity increases CP
concentration and improve digestibility, and consequently increase intake and milk
production. On average, in silage harvest studies (93 diets) marginal efficiency of incremental
CP utilization (153 g milk protein/kg increase in CP intake) was greater than the
corresponding responses to rapeseed feeds (136 and 133) and soybean meal (98)
124
Proceedings of the 5th Nordic Feed Science Conference
Ruminant nutrition
supplementation in the meta-analysis by Huhtanen et al. (2011). However, increased ME
supply is the most likely explanation to increased milk protein yield with earlier harvest.
Intake of ME explained the variation in protein yield much better than CP intake, and in the
bivariate model (ME intake + CP concentration) the effect of CP was non-significant (P =
0.43) and numerically even negative. Although early harvest of silages markedly increase
milk protein yield, the diets based on high CP silages (>140 g/kg DM) were almost as
responsive to increased supplementary protein (rapeseed feeds) as the diets based on low CP
silages (129 vs. 144 g/kg increase in CP intake). It can be concluded that earlier harvest of
silages increase milk yield mainly due to increased ME intake, and positive responses to
supplementary protein can be expected even with highly digestible high CP silages.
Ruminal degradability of forage protein
Determination of ruminal degradability of forage CP has been intensively studied during the
last three decades by in situ incubation technique. More recently, chemical fractionation of
forage CP has been used to estimated ruminal degradability (Cornell system) and
degradability is calculated using different fractional degradation rates for the fractions. The
general conclusion of these studies is that effective ruminal protein degradability (EPD)
increase with increasing buffer / water solubility and CP concentration and decrease with
increasing NDF concentration (decree- sed digestibility) and increasing DM concentration.
This usually increases the contribution of RUP to AAT with advancing maturity of forages.
For example, according to NorFor feed tables AAT/NEL ratio decreases with increased NEL,
i.e. early harvested silages are more limited in AAT relative to NEL than late harvested
silages.
Rinne et al. (2009) calculated silage AAT values for 397 diets either using constant EPD for
all silages or EPD estimated from empirical equations of Yan and Agnew (2004). Their
equations were based on in situ incubations 136 silages. They used DM, CP, NDF, soluble N
and lactic acid/VFA as predictors of silage EPD. The models explained about 80% of
variation in EPD. Interestingly, AAT intake based on constant forage EPD predicted milk
protein yield responses better than AAT intake calculated using EDP values estimated using
Yan and Agnew equations. Interestingly, the coefficients of CP and NDF in the EPD-model
were almost opposite to the coefficients used to correct EPD values for microbial
contamination. Proportion of microbial N in undegraded residues increases with increased
NDF (reduced digestibility) and declining forage CP concentration. It appears that
differences in forage EPD determined by the in situ technique reflect more variation in
microbial contamination than true differences. In addition to microbial contamination, the in
situ technique has many inherent technical problems: escape of soluble NAN from the bags,
loss of feed particles, and lower microbial activity within bags than in rumen digesta,
inappropriate kinetic models. In addition, secondary particle loss can occur when the particles
size decreases during fermentation. Considering the technical difficulties of the in situ
method and the lack of evidence of improved predictions of production responses with feed
specific EPD values using the in situ method for estimating forage EPD can hardly be
justified.
Extent of in-silo fermentation
Both theoretical calculations and experimental evidence suggest that the efficiency of
microbial protein synthesis decrease with increased extent of silage fermentation. This is
because lactic acid and VFA provide less or no energy for rumen microbes. However,
Proceedings of the 5th Nordic Feed Science Conference
125
Ruminant nutrition
discounting fermentable OM for silage acids in calculating forage AAT value did not
improve the predictions of milk protein yield (Rinne et al., 2009), rather vice versa. This can
be due to fermentation of silage lactic acid to propionic acid that the main glucose precursor.
It is possible that increased supply of glucose per unit of intake was greater with extensively
vs. restrictively fermented silages. This can thereby reduce the utilization of amino acids for
glucose production and improve the efficiency of the utilization of absorbed amino acids for
milk protein synthesis. As for earlier harvest of silage, greater milk protein yield in cows fed
restrictively fermented silages is derived from increased feed and ME intake. It may be
concluded that discounting for silage fermentation acids in calculating silage AAT values can
result in greater errors than ignoring the discount for fermentation acids. If discounts are
made then also the effects of increased glucose supply must be taken into account.
Hay vs. silage
In many feed protein evaluation systems, the AAT value of hay is higher than that of silage
when harvested at the same stage of maturity. For example according to NorFor feed tables
AAT/NEL is about 2.5 g/MJ greater for hay than silage at same NEL concentration. This is
both due to discounts of silage fermentation acids in calculating microbial protein and lower
EPD for hay compared to silage when determined by the in situ technique. However,
experimental evidence from milk production studies conducted in SLU (Bertilsson, 1983) and
MTT (e.g. Huhtanen, 1994; Shingfield, 2002) does not give any support for a greater protein
value for hay compared to well-fermented silages from the same sward. In the flow study
with duodenally cannulated cattle (Jaakkola and Huhtanen, 1993), there was no difference in
duodenal NAN flow between diets based on silage or hay, each fed at three levels of contrite
supplements.
N solubility
According to the original model calculating EPD from the in situ kinetic data it was assumed
that immediately disappearing “a-fraction” is completely degraded in the rumen. However,
several experimental techniques have demonstrated that a considerable fraction of soluble
NAN can escape ruminal degradation. Although the degradation rate of soluble NAN
(SNAN) fractions is much faster than insoluble N, much faster passage rate of fluid phase
partly compensates for the difference in degradation rate between soluble and insoluble N
fractions. In addition, feed particles are selectively retained in the rumen, which will increase
ruminal degradability compared with models assuming a single compartment passage kinetic
model. Further, all indigestible CP is included in the insoluble fraction of CP, which also
decreases the difference in the supply of digestible RUP between soluble NAN and insoluble
forage N fractions. In a meta- analysis, silage SNAN fraction had not significant effect on
milk protein yield. SNAN fraction was included in the regression model together AAT intake
calculated using a constant EPD for all silages irrespective of the proportion of soluble N
(Huhtanen et al., 2009). If SNAN had had a specific negative effect on protein yield, the
regression coefficient should have been significantly negative. This indicates that increased
proportion of SNAN in the silage had no negative effect on milk protein yield compared with
insoluble N. In contrast, the proportion of ammonia N had a significant negative effect on
milk protein yield.
It is possible to take into account escape of SNAN in kinetic models, but the methodology to
determine degradation rates of free amino acids, peptides and soluble protein from different
feeds would be a demanding task. Benefits in practical feed evaluation can be questionable.
126
Proceedings of the 5th Nordic Feed Science Conference
Ruminant nutrition
Red clover has much lower proportion of soluble N than grass silage, but the flow of protein
in <38 μm fraction (soluble NAN, particles <<38 μm) was much higher for diets based on red
clover silage compared with grass silage (Huhtanen et al., 2014). In this study in situ method
predicted feed N flow in particle fraction reasonably well, but estimated in situ feed N flow
was poorly correlated with in vivo feed NAN flow.
Red clover
Polyphenol oxidase system decreases proteolysis of red clover both during in-silo
fermentation and in the rumen. Red clover protein is utilized efficiently in the rumen, and the
recovery of incremental protein in the outflow from the rumen is usually high (Vanhatalo et
al., 2009). However, although the NAN flow from the rumen was greater with red clover
compared to grass silages (Dewhurst et al., 2003; Vanhatalo et al., 2009), milk protein yield
did not increase indicating poor utilization of the incremental NAN flow. This can partly be
due to lower intestinal digestibility of red clover protein than grass protein. In Finnish
digestion trials in sheep fed at maintenance faecal CP output was 16 g/kg DM intake greater
for red clover silages compared than for primary growth grass silages. Another reason for the
small production responses to red clover silages can be that ME intake is the limiting factor.
For example, in the study of Vanhatalo et al. (2009) milk protein yield responses were more
closely related to intake of digestible OM than to NAN flow or methionine flow.
Prediction model of milk protein yield
A model of forage factors predicting milk protein yield is presented in Table 1. In addition to
forage factors, the model included linear and quadratic effects of concentrate AAT intake.
Table 1 The best-fit mixed model regression equation of milk protein yield (g/d) to forage variables (RMSE
adjusted for random study effect = 16.0). Adopted from Huhtanen et al. (2010)
Effect
Unit
Intercept
Estimate
Error
P-value
Sda
Response
per SD unit
-310
46.8
<0.001
kg/d
790
66.7
<.0001
CAAT × CAAT
kg/d
-192
38.7
<.0001
Forage DMIc
kg/d
27.7
1.42
<0.001
1.53
42.4
g/kg
0.49
0.069
<0.001
40.9
20.1
g/kg
0.417
0.114
<0.001
22.4
9.3
CAAT
b
d
D-Valued in DM
Forage CP in DM
Ammonia N in total N
4.7
g/kg
-0.217
0.067
0.001
21.5
a
SD = standard deviation; bCAAT = concentrate metabolisable protein; cDMI = dry matter intake; dD-value =
digestible organic matter in DM.
Of forage factors intake and D-value were the most important emphasizing the importance of
ME intake in regulating milk protein yield. The effect of forage CP concentration on milk
protein yield was small (about 4 g per 10 g/kg DM increase in forage CP), although
significant. The key role of energy intake was also demonstrated in meta-analysis by
Huhtanen and Hristov (2009). Only 6–8% of increased supply of AAT (calculated according
to NCR (2001) system) obtained by decreasing protein degradability (constant energy and CP
intake) was recovered as milk protein. This indicates that potential to increase milk protein
yield by manipulating forage CP degradability are minimal, provided that there are no
associated increases in energy intake.
Proceedings of the 5th Nordic Feed Science Conference
127
Ruminant nutrition
Conclusions
There are considerable differences in protein value of forages, but these differences are
mainly associated to intake potential and digestibility of forages. It seems that forage AAT
value is almost entirely related microbial protein (digestibility), whereas differences in
protein fractions have much smaller influence. There is very little experimental evidence
from milk production studies indicating that forage protein has any specific effects on milk
protein yield beyond those related to differences in DM and ME intake. Therefore, accurate
methods/models predicting forage intake potential and digestibility are more important than
complicate models estimating the supply of AAT from undegraded forage protein. Prediction
error of milk protein yield was small even when forage AAT was estimated by a simple
model (microbial AAT from digestible OM – RUP and feed AAT using a constant EPD and
digestibility of RUP) from large datasets (>1000 treatment means) (Huhtanen and
Nousiainen, 2012). This suggests very limited potential for the improvements by more
detailed protein models unless better experimental methods than in situ or chemical
fractionation are developed.
Referenes
Bertilsson, J., 1983). Effects of conservation method and stage of maturity upon the feeding
value .of forages to dairy cows. Swedish University of Agricultural Sciences. PhD thesis,
Report 104.
Dewhurst, R.J., Fisher, W.J., Tweed, J.K.S., & Wilkins, R.J., 2003. Comparison of grass and
legume silages for milk production. 1. Production response with different levels of
concentrate. J. Dairy Sci. 86, 2598–2611.
Huhtanen, P. 1994. Forage influences on milk composition. In A. Fredeen (ed.) Proceedings
of Nova Scotia Forage Conference, Forage: Seeding to Feeding, October 29–30, 1993, p.
144–162.
Huhtanen, P., Bayat, A.R., Krizsan, S.J. & Vanhatalo, A., 2014. Compartmental flux and in
situ methods underestimate total feed nitrogen as judged by the omasal sampling method due
to ignoring soluble feed nitrogen flow. Br. J. Nutr. 111, 535-546.
Huhtanen, P., Hetta, M. & Swensson, C., 2011. Evaluation of canola meal as a protein
supplement for dairy cows: a review and meta-analysis. Canad. J. Anim. Sci. 91, 529–543.
Huhtanen, P. & Hristov, A.N., 2009. A meta-analysis of the effects of protein concentration
and degradability on milk protein yield and milk N efficiency in dairy cows. J. Dairy Sci. 92,
3222–3232.
Huhtanen, P. & Nousiainen, J., 2012. Production responses of lactating dairy cows fed silagebased diets to changes in nutrient supply. Livest. Prod. Sci. 148, 146–158.
Huhtanen, P., Rinne, M. & Nousiainen, J., 2008. Effects of silage soluble N components on
metabolizable protein concentration: a meta-analysis of dairy cow production experiments. J.
Dairy Sci. 91, 1150–1158.
Huhtanen, P., Sudekum, K-H., Nousiainen, J. & Shingfield, K., 2010. In: Schnyder H. et al.
(eds). Grassland in a Changing World. Grassl. Sci. Europe 15, 379–400.
128
Proceedings of the 5th Nordic Feed Science Conference
Ruminant nutrition
Jaakkola, S. & Huhtanen, P., 1993. The effects of the forage preservation method and the
proportion of concentrate on nitrogen digestion and rumen fermentation in cattle. Grass and
Forage Sci. 48, 146–154.
Rinne M., Huhtanen, P. & Nousiainen, J., 2009. Effects of silage effective protein
degradability and fermentation acids on metabolizable protein concentration: a meta-analysis
of dairy cow production experiments. J. Dairy Sci. 92, 1633–1642.
Shingfield, K.J., Jaakkola, S. & Huhtanen, P., 2002. Effect of forage conservation method,
concentrate level and propylene glycol on intake, feeding behavior and milk production of
dairy cows. Anim. Sci. 74, 383–397.
Shingfield, K.J., Jaakkola, S. & Huhtanen, P., 2001. Effects of level of nitrogen fertilizer
application and various nitrogenous supplements on milk production and nitrogen utilization
of dairy cows given grass silage-based diets. Anim. Sci. 73, 541–554.
Vanhatalo, A., Kuoppala, K., Ahvenjärvi, S. & Rinne, M., 2009. Effects of feeding grass or
red clover silage cut at two maturity stages in dairy cows. 1. Nitrogen metabolism and supply
of amino acids. J. Dairy Sci. 92, 5620–5633.
Vanhatalo, A. & Toivonen, V., 1993. Influence of cutting time and N fertilization of grass
silage and level of grass fertilized with different levels of N. Proceedings of the 10th
International Conference on Silage Research, Dublin, Ireland, Sept. 6th–8th 1993. pp. 168–
169.
Yan T., & Agnew, R. E., 2004. Prediction of nutritive value of grass silages: II. Degradability
of nitrogen and dry matter using digestibility, chemical composition and fermentation data. J.
Anim. Sci. 82, 1380–1391.
Proceedings of the 5th Nordic Feed Science Conference
129
Horses
The new protein evaluation of horse feeds in Germany
K.-H. Südekum
on behalf of the Committee for Requirement Standards of the Society of Nutrition Physiology
Institute of Animal Science, University of Bonn, Endenicher Allee 15, 53115 Bonn, Germany
Correspondence: [email protected]
Background
In equines, amino acids (AA) released through auto-enzymatic protein digestion can be
absorbed from the small intestine, whereas AA from microbial protein production and breakdown in the hindgut are not absorbed (Schubert, 1992) or absorbed in small amounts – from
studies using labeled amino acids, up to 12% of the plasma amino acids originated from the
large intestine (McMeniman et al., 1987). Woodward et al. (2010), based on AA transporter
mRNA abundance in the large intestine, have concluded that “… results indicate that the
large intestine might contribute to both cationic and neutral AA uptake and absorption”.
However, the potential contribution of the large intestine to AA supply of horses still needs to
be quantified.
Therefore, a concept of protein evaluation that allows estimating the extent of auto-enzymatic
digestion appears to be more appropriate and should result in more reliable data on the
protein or AA supply to the horse than current systems that base on crude protein (CP) or
digestible CP (DCP), the latter sometimes being estimated from CP (Pálson, 1973). In horses,
the experimental data base on small intestinal digestibility (siD) of protein is too small for
establishing a system based on small intestinally digestible CP (sidCP) or AA (sidAA), which
would be in accordance with the ‘ileal’ or ‘pre-caecal’ digestibility variables used in other
non-ruminant species such as pigs. While constructing a database for sidCP would have been
possible, and this could have been extended by including data from studies using the mobile
nylon bag technique (Martin-Rosset et al., 2012), only two (2!) studies have been published
on measured sidAA (Almeida et al., 1998; Coleman et al., 2000), which has made it
impossible to develop a consistent system of sidCP and sidAA similar to those in pigs that
rely on measured data using strictly standardized methodology. As a consequence of this
current status, the following idea was developed.
Based on the CP partitioning of the ‘Cornell Net Carbohydrate and Protein System’ for cattle
using standardisation and recommendations published by Licitra et al. (1996), two CP
fractions, namely (i) neutral detergent insoluble CP (NDICP; insoluble protein) and (ii) the
corresponding ND-soluble CP (NDSCP; soluble protein; CP - NDICP) were used. The idea
behind this fractionation is that CP bound to cell-wall material and, thus, recovered in the
ND-insoluble residue cannot be broken down by auto-enzymatic digestion, whereas the
soluble protein can be digested by body’s own enzymes. As both cattle and horses are
herbivores and may consume very similar diets and diet ingredients, data on NDICP in forage
and concentrates available for cattle may also be used for constructing a horse database.
The aim of the present study was to investigate whether the suggested concept of protein
evaluation on the basis of ND-soluble versus ND-insoluble CP is applicable to horse feeds.
The following sections briefly summarizes work conducted by the Committee for
Requirement Standards (AfBN, ‘Ausschuss für Bedarfsnormen’) of the Society of Nutrition
Physiology (GfE, Gesellschaft für Ernährungsphysiologie) in Germany and is largely based
on contributions of horse experts, namely (alphabetical order) Manfred Coenen, Ellen
130
Proceedings of the 5th Nordic Feed Science Conference
Horses
Kienzle and Annette Zeyner. The outcome of this work will be published soon in a volume of
the GfE series “Recommendations for the supply of energy and nutrients” (GfE, in press).
Materials and Methods
Neutral detergent insoluble CP and NDSCP were analysed and defined as given above, i.e.,
NDSCP = CP - NDICP. For silages, a correction was made for ammonia-N as follows:
DSCP = CP - NDICP - NH3-N · 6.25. Data on NDICP of forages and concentrates were taken
from published work on ruminant feedstuffs (for references, see GfE, in press). A literature
survey was then conducted to investigate if the above fractionation scheme can be applied to
estimate DCP and siD of CP and AA, respectively.
Results and Discussion
The survey included the following steps and major results (references are given in GfE, in
press and can be obtained upon request from the author).
1. A correlation exists between NDSCP of feeds and siDCP in the horse and, likewise,
between NDICP and the CP that is indigestible in the small intestine (R = 0.719, p < 0.01; N
= 24).
2. A strong linear correlation exists between the daily intake of siDCP and NDSCP (r =
0.832, p < 0.003). Although based on only few experimental data, siDCP was measured on a
range of rations reflecting typical horse diets. The slope of the regression between intakes of
siDCP and NDSCP indicates that 0.9 of NDSCP intake is digested (and the AA are being
absorbed) within the small intestine. Therefore, it is assumed that:
siDCP = 0.9 · NDSCP.
Moreover, the AA profiles of both, NDSCP and NDICP, appear to be similar within a given
feed (Rebolé et al., 2001; Tedeschi et al., 2001), such that the NDSCP/NDICP schedule can
be translated to a NDSAA/NDIAA schedule and the above equation can also be used to
estimate siDAA as follows:
siDAA = 0.9 · NDSAA.
No data could be found on the fate of supplemented free AA in horses. It was assumed,
therefore, that free AA are completely absorbed from the proximal small intestine and thus,
siD of supplemented free AA is 100%.
Feedstuffs and mixed diets typically fed to horses are currently being evaluated in terms of
supply of siDCP and siDAA. In a final step, the match between supply and requirements for
CP and AA, also expressed as siDCP and siDAA, has to be assessed.
Conclusions
Neutral detergent-soluble crude protein can be easily calculated as the difference of analysed
feed fractions, i.e. CP minus NDICP, and represents the CP fraction which can potentially be
digested in the small intestine. Based on very limited evidence from horse studies, data on
other non-ruminant species (i.e., pigs) and theoretical considerations, the siD of this fraction
was estimated to be 90%. Because of the similar AA profiles of NDSCP and NDICP the
concept can be translated to estimate siDAA from feed chemistry data without requiring
animal experimenttation. The GfE (in press) recommends to include siD of lysine,
(methionine + cysteine) and threonine in table values for horse feeds. The new system,
Proceedings of the 5th Nordic Feed Science Conference
131
Horses
although based on very limited direct experimental evidence, more closely resembles the
conditions in the gastro-intestinal tract of horses than the DCP systems currently used.
Moreover, the new system can be easily extended and adopted to new data.
References
Almeida, F.C., Valadares Filho, S.C., Queiroz, A.C., Leão M.I., Silva, J.F.C. & Cecon, P.R.,
1998. Digestibilidade aparente e verdadeira pré-cecal e total da proteína em dietas com
diferentes níveis protéicos em eqüinos. R. Bras. Zootec. 27, 521-529.
Coleman R. J., Matthison G.W., Hardin R.T. & Milligan J.P., 2000. Effect of dietary forage
and protein concentration on total tract, prececal and post-ileal protein and lysine
digestibilities of forage based diets fed to mature ponies. 17th Equine Nutr. Physiol. Symp.
Lexington, Kentucky, USA, pp. 461-463.
GfE (Gesellschaft für Ernährungsphysiologie), in press. Empfehlungen zur Energie- und
Nährstoffversorgung der Pferde. DLG-Verlag, Frankfurt/Main (in German).
Licitra, G., Hernandez, T.M. & Van Soest, P.J., 1996. Standardization of procedures for
nitrogen fractions of ruminant feeds. Anim. Feed Sci. Technol. 57, 347-358.
Martin-Rosset, W., Macheboeuf, D., Poncet, C. & Jestin, M., 2012. Nitrogen digestion of a
large range of hays by mobile nylon bag technique (MNBT) in horses. In: Saastamoinen, M.,
Fradinho, M.J., Santos, A.S., Miraglia, N. (Eds.), Forages and Grazing in Horse Nutrition.
EAAP Publ. No. 132, Wageningen Academic Publishers, Wageningen, The Netherlands, pp.
109-120.
McMeniman, N. P., Elliott, R., Groenendyk, S. & Dowsett, K. F., 1987, Synthesis and
absorption of cysteine from the hindgut of the horse. Equine Vet. J. 19, 192-194.
Pálson, T., 1973. Bestämning av råproteinets smältbarhet i vallfoder (Determination of crude
protein digestibility in forage). Department of Animal Nutrition and Management, Swedish
University of Agricultural Sciences. Uppsala, Sweden.
Rebolè, A., Treviño, J., Caballero, R. & Alzueta, C.,2001. Effect of maturity on the amino
acid profiles of total and nitrogen fractions in common vetch forage. J. Sci. Food Agric. 81,
455-461.
Schubert, R., 1992. Verwertung von 15N-Harnstoff für die intestinale Synthese von
Bakterienprotein und für die Milchbildung bei Ponystuten. Pferdeheilkde. 43 (Special Issue),
137-139 (in German).
Tedeschi, L.O., Pell, A.N., Fox, D.G. & Llames, C.R., 2001. The amino acid profiles of the
whole plant and of four plant residues from temperate and tropical forages. J. Anim. Sci. 79,
525-532.
Woodward, A.D., Holcombe, S.J., Steibel, J.P., Staniar, W.B., Colvin, C. & Trottier, N.L.,
2010. Cationic and neutral amino acid transporter transcript abundances are differentially
expressed in the equine intestinal tract. J. Anim. Sci. 88, 1028-1033.
132
Proceedings of the 5th Nordic Feed Science Conference
Horses
Effects of crude protein content in forage-only diets fed to horses
A. Jansson1 & S. Ringmark1
1
Swedish University of Agricultural Sciences (SLU), Department of Animal Nutrition &
Management, Monogastric Division, Almas allé 8D , 750 07, Uppsala, Sweden
Correspondence: [email protected]
Introduction
The advantages of feeding diets dominated by forage to horses are many, i.e. there appears to
be a decreased risk of several health problems such as colic, rhabdomyolysis, gastric ulcers
and stereotypic behaviors compared to traditional high concentrate, starch rich diets. Forages
with a great variation in energy content can be produced by harvesting grasses at different
stages of maturity. However, the content of crude protein (CP) may show an even bigger
variation and both under and over feeding with CP is likely to be common. There are not
many studies performed on effects of protein intake on the metabolism of horses, especially
not in exercising horses (Jansson and Harris, 2013) and in horses fed forage-only diets. The
aim of the present study is to highlight some of the recent studies performed on these
subjects.
Effects of protein intake on digestion and hindgut environment
Being a grass eater, horses are adapted to a large variation in crude protein intake. The
content of CP in forages and pastures varies with the maturity of the plants, plant species, dry
periods/rainy periods, the number of animals grazing/ha, soil conditions and fertilization. The
most common deficiency of the diet in free ranging herbivores is a protein/energy deficiency
(Duncan, 1992) but in adult, non-productive (pleasure) horses, short pastures are likely to
provide excessive amounts of CP.
Protein is mainly digested and absorbed in the small intestine. The digestibility of forage
protein is affected by the maturity of the grasses. In mature, low CP forages (~8 % CP of dry
matter (DM)) digestibility may be as low as 43 % and in forages with higher CP contents
higher than 65 % (NRC, 1989). Excessive nitrogen is mainly excreted in the urine
(Connysson et al., 2006). However, part of the surplus nitrogen enters the hind gut and affects
not only the fecal nitrogen excretion (Connysson et al., 2006) but also the hind gut
environment and microflora. We have observed small but significant drops in colon and fecal
pH and DM content on high CP forage-only diets (Connysson et al., 2006; Muhonen et al.,
2008a) and also an increase in total VFA concentrations in colon (Muhonen et al., 2008a).
Excessive CP intake also seems to increase heat production and evaporative losses
(Muhonen, 2008b), probably as a result of the increased production of urea. During hot and
humid conditions this might be a limitation to exercise performance. The increased urea
production may also affect acid base balance during exercise and the importance of this for
performance remains to be proven, since results so far are not consistent (Graham-Thiers et
al., 2001; Connysson et al., 2006).
The use of fecal samples to determine crude protein intake
Horses are selective grazers (Duncan, 1992) and to get a representative sample of what they
consume at pasture may therefore be complicated. If protein intake could be monitored with
reasonable precision in grazing horses the possibility to adjust deficiencies and make
necessary supplementation would be improved. Fecal samples have earlier been used to
estimate CP in diet of feral horses (Duncan, 1992; Salter and Hudson, 1979). In these studies,
Proceedings of the 5th Nordic Feed Science Conference
133
Horses
seasonal variations in both CP and fiber content have been observed with highest fiber and
lowest CP contents during wintertime. In a pilot study, we investigated the possibility to use
fecal samples as a rough marker for CP intake in horses on forage-only diets. Forage and
fecal samples were collected from stabled horses observed in three studies (A, B and C).
Study A included 11 adult sedentary horses (11 samples) of various breeds from different
stables (forage CP range: 7.8-10.6 % of DM). Study B included 16 Standardbred horses
sampled once as 1-year olds and twice as 2- and 3-year olds respectively, while fed four
different grass forage batches complemented daily with 250-1000 g of a lucerne product (65
samples, forage CP range (including lucerne): 9.8-16.4 % of DM). Study C included six adult
trained horses on three grass forage batches (18 samples, forage CP range: 9.1-13.4 % of
DM, Ringmark and Jansson, 2013). Feed and feces samples were collected in the same day
except in study B where feed samples were collected for three consecutive days and mixed
before analysis and the fecal samples used were from day 2 or 3.
CP in feces varied between 6.6 and 18.1 % of DM. There was a significant, but not very
strong, correlation (Pearson´s correlation coefficient) between CP in feed and CP in feces (R
=0.60, P <0.0001, figure 1). The correlation was not improved if data was separated into
growing (< 3 years) and adult horses. The effect of CP intake on fecal CP content was also
tested in a mixed model including effect of study and repeated measurements on individuals
which showed an effect of CP in feed on CP in feces (P <0.001). Additionally, observations
were divided in two levels based on the CP content in feed DM (medium (n=52): < 13% CP
and high (n=47): ≥ 13% CP) and the effect on fecal CP content was analyzed with a general
linear model including effect of individual. This resulted in significantly higher content of CP
in feces in the high level than in the medium level (11.3 ± 0.3 and 10.1 ± 0.2 % of DM,
respectively, P =0.0001). However, since there were no observations on forages with CP
content lower than 7.8 % of DM, still no recommendations could be made for how this
method could be used to assess CP intakes below requirements. The nutritive quality of
pasture and the body condition of the horses should therefore always be considered primarily.
Figure 1 The relation between crude protein content in feed and in feces (% of dry matter). Y=0.6709x+5.0258.
Crude protein content may affect insulin response
Forage-only diets result in lower basal and post-prandial plasma insulin concentrations than
starch-rich concentrate based diets (Connysson et al., 2010; Jansson and Lindberg, 2012).
However, it has also been shown that the content of water soluble carbohydrates (WSC) in
134
Proceedings of the 5th Nordic Feed Science Conference
Horses
forage affects the insulin response to feeding (Borgia et al., 2011; Ragnarsson and Jansson,
2011). In both horse and man it is also known that protein, and especially the amino acid
leucine, may induce an increased insulin response (Cleator et al.,1975; Urschel et al., 2010)
and in humans, an additional effect on insulin response with the simultaneous ingestion of
amino acids and glucose compared with the ingestion of glucose alone have been observed
(van Loon et al., 2000). The plasma insulin level is likely to affect both sedentary and athletic
horses. In horses suffering from equine metabolic syndrome, high insulin levels might
increase the risk for laminitis and in athletic horses insulin levels will affect the metabolic
response to exercise and possible the performance. To assess the insulin response to forageonly diets with varying crude protein content both at rest and after exercise, a study on six
adult conditioned Standardbred horses was recently conducted (Ringmark and Jansson,
2013). Two levels of CP intake was studied, ≤180 g CP/100 kg body weight (BW) and >180
g CP/100 kg BW. The intake of CP/100 kg BW was significantly higher (P <0.05) in the high
CP-intake group than in the low CP- intake group, but there was no difference in WSC
(excluding fructans) intake. There was a tendency (P=0.08) for higher plasma insulin
response after feeding in the high CP group compared to the low group but no difference
between groups was observed after feeding when horses had been exercised within one hour.
An ANOVA model including horse, CP and WSC (excluding fructans) intake explained 95 %
of the variation in plasma insulin whereas a model including horse and WSC intake only
explained 87 % of the variation in plasma insulin. Although this study had a small number of
observations, it indicates that insulin response to feeding forage can be elevated by a high CP
content. This is also in accordance with previous observations in horses fed readily available
carbohydrates together with protein or amino acids (Stull and Rodiek et al., 1966; Rérat et al.,
1985; van Loon et al., 2000). However, the importance of CP intake for insulin response does
not seem to be as central post exercise as before, at least not at the CP intakes level studied in
the present experiment.
High glycogen levels on a high CP diet
Low plasma insulin levels could be an advantage for the athletic horse since low levels
promotes increased plasma glucose levels and fat availability during exercise (Pagan and
Harris, 1999), factors likely to improve performance. However, low levels could also be a
limitation since insulin may stimulate replenishment of muscle glycogen and fat deposition.
However, we have shown that horses fed a high CP forage-only diet will have very high
muscle glycogen contents (Essén-Gustavsson et al. 2010), comparable to horses fed high
concentrate diets, and also that replenishment may be improved compared to a more
moderate CP intake (Figure 2). Together with the observations made by Ringmark and
Jansson (2013), this implies that high insulin levels are not critical to the immediate post
exercise glycogen replenishment. However, high muscle glycogen contents on protein rich
diets are also in accordance with observations in rats, and it has been suggested that intake of
branched amino acids stimulate glycogen synthesis through insulin-independent mechanisms
(Nishitani et al., 2002; Morifuji et al., 2010).
Conclusions
Protein intake may show a great variation in horses and intake level will affect the digestive
system and metabolism at rest as well in connection with exercise. However, further studies
are needed to better understand the importance of the effects and how protein intake can be
monitored in horses on pasture or mixed forage/pasture diets.
Proceedings of the 5th Nordic Feed Science Conference
135
Horses
Figure 2 Muscle glycogen content (mmol/kg dry weight) before, immediately after and 90 min after an exercise
test in horses fed a forage-only diet providing recommended (black bars, 12.5 % of DM) and high crude protein
intakes (grey bars, 16.6 % of DM). The glycogen content was higher on the high CP-diet and the there was a
significant (P<0.05) decrease after exercise in both diets and a tendency (P=0.09) to an increase after 90 min on
the high CP-diet but not on the recommended diet. After Essén-Gustavsson et al. (2010).
References
Borgia, L., Valberg, S., McCue, M., Watts, K. & Pagan, J., 2011. Glycaemic and
insulinaemic responses to feeding hay with different non-structural carbohydrate content in
control and polysaccharide storage myopathy-affected horses. J. Anim. Physiol. Anim. Nutr.
95, 798-807.
Connysson, M., Muhonen, S., Lindberg, J. E., Essén-Gustavsson, B., Nyman, G., Nostell, K.
& Jansson, A. 2006. Effects on exercise response, fluid and acid-base balance of protein
intake from forage-only diets in Standardbred horses. Equine Vet. J. Suppl. 36, 648 - 653.
Connysson, M., Essen-Gustavsson, B. & Jansson, A., 2010. Effects of feed deprivation on
Standardbred horses fed a forage-only diet and a 50:50 forage-oats diet. Equine Vet. J. 42,
335-340.
Duncan, P. 1992. Horses and Grasses. The Nutritional Ecology of Equids and Their Impact
on the Camargue. Ecological studies 87, Springer-Verlag, New York.
Essén-Gustavsson B., Connysson, M. & Jansson, A., 2010. Effects of crude protein intake
from forage-only diets on muscle amino acids and glycogen levels in horses in training.
Equine Vet. J. 42, 341-346.
Graham-Thiers, P.M. , Kronfeld, D.S., Kline, K.A., Sklan, D.J. 2001. Dietary protein
restriction and fat supplementation diminish the acidogenic effect of exercise during repeated
sprints in horses. J. Nutr. 131(7), 1959-1964.
Jansson, A. & Harris, P.A. 2013. A bibliometric review on nutrition of the exercising horse
from 1970 to 2010. Comp. Exercise. Physiol. Vol. 9, Number 3 - 4.
Jansson, A. & Lindberg, J.E., 2012. A forage-only diet alters the metabolic response of
horses in training. Anim. 6, 1939-1946.
136
Proceedings of the 5th Nordic Feed Science Conference
Horses
Muhonen, S.; Connysson, M.; Lindberg, J. E.; Julliand, V.; Bertilsson, J.; Jansson, A.2008a.
Effects of crude protein intake from grass silage-only diets on the equine colon ecosystem
after an abrupt feed change. J. Anim. Sci. 86, 3465-3472.
Muhonen, S. 2008b. Metabolism and hindgut ecosystem in forage fed sedentary and athletic
horses. Acta Universitatis Agriculturae Sueciae. Doctoral Thesis No. 2008:68. Swedish
University of Agricultural Sciences, Uppsala, Sweden.
Morifuji, M., Kanda, A., Koga, J., Kawanaka, K. & Higuchi, M., 2010. Post-exercise
carbohydrate plus whey protein hydrolysates supplementation increases skeletal muscle
glycogen level in rats. Amino Acids 38, 1109-1115.
Nishitani, S., Matsumura, T., Fujitani, S., Sonaka, I., Miura, Y. & Yagasaki, K., 2002.
Leucine promotes glucose uptake in skeletal muscles of rats. Biochem.Biophysical Res.
Comm. 299, 693-696.
NRC.1989. Nutrient Requirements of Horses (5th Revised Ed.). Nutrient Requirements of
Domestic Animals. National Research Council. National Academy Press, Washington, DC.
Pagan, J. & Harris, P., 1999. The effects of timing and amount of forage and grain on
exercise response in thoroughbred horses. Equine Vet. J. Suppl. 30, 451-457.
Ragnarsson S & Jansson A. 2011. A comparison of grass haylage digestibility and metabolic
plasma profile in Icelandic and Standardbred horses. J. Anim. Physiol. Anim. Nutr. 95, 3,
273-279.
Rérat, A., Chayvialle, J.A., Kande, J., Vaissade, P., Vaugelade, P. & Bourrier, T., 1985.
Metabolic and hormonal effects of test meals with various protein contents in pigs. Canadian
J. Physiol. Pharmacol. 63, 1547-1559.
Ringmark, S & Jansson, A., 2013. Insulin response to feeding forage with varying crude
protein and amino acid content in horses at rest and after exercise. Comp. Exercise Physiol.
Vol. 9, Number 3 - 4.
Salter, R.E. & Hudson, R. J., 1979. Feeding Ecology of feral horses in western Alberta. J.
Range Managm., 32, 221-225.
Stull, C.L. & Rodiek, A.V., 1988. Response of blood glucose, insulin and cortisol
concentrations to common equine diets. J. Nutr. 118, 206-213.
Urschel, K.L., Geor, R.J., Waterfall, H.L., Shoveller, A.K. & McCutcheon, L.J., 2010.
Effects of leucine or whey protein addition to an oral glucose solution on serum insulin,
plasma glucose and plasma amino acid responses in horses at rest and following exercise.
Equine Vet. J. 42, 347-354.
van Loon, L.J.C., Saris, W.H.M., Verhagen, H. & Wagenmakers, A.J.M., 2000. Plasma
insulin responses after ingestion of different amino acid or protein mixtures with
carbohydrate. American J. Clinical Nutr. 72, 96-105.
Proceedings of the 5th Nordic Feed Science Conference
137
Methods and miscellaneous
Simple Excel VBA application for Monte Carlo simulation with ad hoc made
spreadsheet models
T. Eriksson
Department of Animal Nutrition and Management, Swedish University Agricultural
Sciences, Kungsängen Research Center, SE-753 23 Uppsala, Sweden.
Correspondence: [email protected]
Introduction
Researchers in animal science as well as in many other disciplines use spreadsheet models in
a multitude of ways. There are large, well organized models created with considerable efforts.
But, the majority of models are ad hoc made applications quickly put together when checking
up some idea that struck the researcher examining the outcome of an experiment or reviewing
a manuscript. These models may use only a handful of cells, but sometimes evolve into huge
poorly organized “spaghetti models” as more factors are taken into account.
A typical issue when examining complex relationships involving multiple factors is what
effect they may have on the variation of the response variables. This could be tested by
Monte Carlo simulation, where, for instance, multiple model parameters are allowed to vary
according to a chosen distribution. The outcome of thousands of simulations generates a
distribution in response variables which could be regarded as an extended sensitivity analysis.
This paper describes a simple spreadsheet application in Microsoft Excel that easily can be
connected to existing ad hoc spreadsheet models for generating Monte Carlo simulations. The
application is free for download and use.
The application
The application (APP) was made in Excel 2010 and consists of a worksheet and a short VBA
code (macro) operated by a button on the worksheet. The VBA code allows simulations to be
run automatically any chosen number of times (thousands). The actual value of each
parameter in each simulation iteration as well as the corresponding values of response
variables are logged and forms the outdataset from the simulation. The worksheet could
either be inserted into the Excel file containing the existing model or, the model could be
copied into the file with the APP. In both cases, the file must be saved as a macro enabled
workbook.
The APP was made as transparent as possible, to facilitate understanding of the construction
and make it easier for advanced users to modify it. The present worksheet has seven rows
with constants or formulae (Figure 1), of which three require user activity for each new
simulation setup. The other rows could be changed by expanding or diminishing the range of
cells in use. The worksheet in the example is set up for a maximum of 30 model parameters
and storage of these and a maximum of 20 model output variables.
Random coefficients are generated in Rows 1 and 2. Row 1 has a range of cells holding
constants. In the example, there are 800 cells with values following a normal distribution with
mean = 0 and s = 1, i. e., the values vary from approx. -3 to 3. The values were generated by
Excel Analysis ToolPak but could also be imported from other sources or replaced by data
following some other distribution if considered more suitable. Row 2 contains
RANDBETWEEN and OFFSET formulas to retrieve a value from a randomly chosen cell in
the range of normally distributed constants in Row 1. The value is replaced at each
138
Proceedings of the 5th Nordic Feed Science Conference
Methods and miscellaneous
recalculation of the worksheet, i.e. at any manual change to the worksheet, when F9 key is
pressed or by ‘Calculate’ in a macro.
Row 3 is for user input of starting values for model parameters to be varied in the
simulations. In the example, the average value for each parameter is used. The standard
deviation for each parameter to be varied is entered in Row 4.
In Row 5, the random coefficients generated in the first step are multiplied with user entered
standard deviations. This gives the absolute deviation for each parameter in the current
simulation. In Row 6, all values are checked for non-negativity and added to the starting
values to yield parameter values to be used in the simulation. If negative, a zero value is used.
Row 7 contains both current values of the parameters and model variable output data of
interest, i.e. result of the simulation. Variable values are linked with Row 6 and with the
existing model. Output data from the existing model are linked to cells in Row 7 to the right
of the parameters. The row is logged below in a designated cell range after each simulation.
Random coefficient generation
Row 1 Cells populated with standardized distribution (In example normal distribution with x = 0; s = 1)
Row 2 Coefficients retrieved from randomly chosen cells containing the distribution
Input of values and known/assumed variation for model parameters of interest
Labels for model parameters
Row 3 Original values for model parameters that should vary in the simulation
Row 4 Known/assumed standard deviation for each varying model parameter
Calculation of parameter value to be used in the simulation
Row 5 Random coefficients are multiplied with the known/assumed parameter standard deviations
Row 6 Results are added to the original values and checked for non-negativety
-2.7107 ……….…''
-0.273 ……….…''
Ash, g/kg DM ……….…''
22.9 ……….…''
2.0
-0.5 ……….…''
22.4 ……….…''
Input in model of parameters with simulated variation and retrieval of results
Parameter values to be used are linked from Row 6 and further to the formulas in the model.
Outdata cells of interest from the model are linked to the same row.
Row 7 The row is after each recalculation copied to a database.
22.4 ……….…''
Figure 1 Structure of the spreadsheet for generating Monte Carlo simulations from existing Excel models. Data
continue to the right in as many columns as needed. User activity is required for entering parameter values and
standard deviations in Rows 3 and 4, respectively and for linking cells to Row 7. Row 7 is logged after each
simulation.
An example
The example chosen is the summative effect of intra-laboratory analytical variation of
laboratory estimated metabolizable energy (ME) content of feeds and resulting ME intake of
a dairy cow diet. Proximate analysis of concentrates with a set of tabulated coefficients for
digestibilty and metabolizability has been used in Sweden since 1967. Together with forage
ME estimation from in vitro organic matter digestibility, it has formed the Swedish system
for ruminant ME calculation from the late 1970’s and onwards and has only recently been
partly replaced by the Norfor model.
A diet with barley, rapeseed cake and grass silage was used. Initially, variation in the ME
concentration of barley was investigated and in the next, total ME intake from the diet and
the ME allowable production of energy corrected milk (ECM) according to Spörndly (2003)
Proceedings of the 5th Nordic Feed Science Conference
139
Methods and miscellaneous
was simulated. Table 1 shows analytical values together with calculated ME concentration of
the feeds and standard deviation for repeated analyses of control samples at the Kungsängen
Research Centre laboratory (Verner et al., 2012).
Table 2 Metabolizable energy calculation from proximate analysis of concentrates and in vitro organic matter
digestibility of forage for the feeds used in the simulation examples. Intra-laboratory standard deviations
(reproducibility over time for the Kungsängen Research Centre laboratory) that were used in the simulations are
included
Barley
Rapeseed
cake
Grass
silage
Reproducibility
(SD between repeated analyses)
Ash, g/kg DM
28
66
60
2
Crude protein, g/kg DM
122
339
0.75
Crude fat, g/kg DM
27
169
1.10
62
137
4.0
761
289
Item
Proximate analysis
Crude fiber, g/kg DM
Nitrogen free extractives, g/kg DM
a
Forage OMD and ME calculation
In vitro OMD of forage, g/kg OM
a
873
MJ ME/ kg DM
13.16
15.46
Calculated as a residual: 1000-ash-crude protein-crude fiber
9.2
11.34
Table 2 Total ME intake and production potential of energy corrected milk (ECM) for the diet in the simulation
example
Item
Barley
Rapeseed
cake
Grass
silage
MJ ME/ kg DM
13.16
15.46
11.34
Intake, kg DM/d
6
2
12
Total diet
20
Total intake, MJ ME/d
246
ME allowable ECM, kg/d
34.5
Results
Simulations were run on a standard laptop computer with 2.1 GHz processor and 8 GB RAM
memory, which required 150 seconds for 10000 simulations of barley ME concentration with
four parameters. For the total diet, ten parameters were used and the simulation required 165
seconds for 10000 simulations. Correlations among parameters were in absolute values
< 0.03. The range of ME values for barley was about 0.8 MJ/kg DM (Figure 2). The total diet
ranged from 239 to 252 MJ ME/d with a standard deviation of 1.8 MJ. The estimated
resulting ECM production supplied (Figure 3) varied over a range of 2.35 kg/d as an effect of
laboratory analytical variation.
140
Proceedings of the 5th Nordic Feed Science Conference
Methods and miscellaneous
2500
Count
2000
1500
1000
13.53
13.47
13.41
13.35
13.29
13.22
13.16
13.10
13.04
12.98
12.92
12.86
12.79
0
12.73
500
MJ ME/kg DM
Figure 2 Metabolizable energy (ME) concentration from 10000 simulations with the example barley. All
variation is for the result of known variation in laboratory analytical data. Range 12.72 – 13.54, s 0.123. Class
width is 0.5 s.
2500
Count
2000
1500
1000
500
35.60
35.43
35.27
35.11
34.94
34.78
34.62
34.45
34.29
34.12
33.96
33.80
33.63
33.47
33.31
0
ME allowable ECM, kg/d
Figure 3 Energy corrected milk (ECM) production as allowed by metabolizable energy (ME) intake from 10000
simulations with the example diet. All variation is due to the known variation for laboratory analyses of the
three feeds in the diet. Range 33.29 – 35.64, s 0.327. Class width is 0.5 s.
Discussion
The simulations displayed the possible variation in outcome when four or ten parameters are
allowed to vary independently following normal distribution. It would have been possible to
let a much larger number of parameters vary if feasible. In this specific case, the assumption
of no covariance between variables is probably correct, because it deals with random
analytical variation at a laboratory. However, when applying this approach to other models,
Proceedings of the 5th Nordic Feed Science Conference
141
Methods and miscellaneous
the possibility of covariance between variables should always be taken into account. Consider
for instance a ruminant nutrition model, where there most likely exists covariance between
ruminal digestibility and small intestinal digestibility of a feed component.
The random generator in Excel has occasionally been criticized (McCullough, 2008), but the
absence of correlation between variables does not suggest systematic errors. It would be
possible to extend the population of coefficients for sampling and also to perform repeated
recalculations in each simulation step, although simulation time would increase.
The situation simulated here has likely occurred in many dairy cow experiments and would
have had minor consequences for result interpretation if: 1) the experiment was of a changeover design and random analytical differences would just add to period effects or 2) feed
samples from the entire experiment were analysed in the same batch. However, if none of
these conditions are fulfilled, the repeatability variation at a research laboratory could cause
differences of the magnitude, shown above.
Analytical variation of a control sample will not reflect what the variation will be in all
sample types and it is likely that variation actually is larger in some cases. Reproducibility
among research laboratories or commercial laboratories is generally greater and if such data
is mixed within an experiment, variation would most likely increase considerably.
Conclusion
A relatively simple Excel APP can be used for investigating the outcome when multiple
parameters are varied independently in simulations. They can easily be linked to existing
Excel models by the user.
References
McCullough, B.D. 2008. Microsoft Excel’s ‘Not The Wichmann–Hill’ random number
generators. Comp. Stat. Data Anal. 52, 4587–4593.
Spörndly, R. 2003.(Ed) Fodertabeller för idisslare (in Swedish). Department of animal
nutrition and management. Swedish Universtity of Agricultural Sciences. SLU/HUV Report
257, pp, 7-15.
Verner, E., Ericson, B. & Udén, P., 2012. The variation of control samples in the analyses of
crude protein, neutral detergent fibre, in vitro organic matter digestibility and crude fat. 3rd
Nordic Feed Science Conference, Uppsala, 28-29 of June 2012. SLU/HUV Report 280, p 2730.
The Excel APP is downloadable at: www.slu.se/en/torsten-eriksson
142
Proceedings of the 5th Nordic Feed Science Conference
Methods and miscellaneous
Comparison between individual feeds and total diets on predicted methane production
from in vitro simulations
M. Ramin1, M. Vaga1, E. H. Cabezas-Garcia1 & E. Detmann2
1
Department of Agricultural Research for Northern Sweden, Swedish University of
Agricultural Sciences, SE-901 83 Umeå, Sweden
2
Departamento de Zootecnia, Universidade Federal de Viçosa, Viçosa, Minas Gerais, Brazil
Correspondence: [email protected]
Introduction
Methane production is a major problem in ruminant production system as it represents a
significant energy loss from the diet i.e., 2–15% of gross energy intake (Johnson and
Johnson, 1995). Many factors influence methane production in ruminants such as
digestibility, type of feed and dry matter intake (Ramin and Huhtanen, 2013). In production
trials, total mixed ration feeding system is applied and occasionally methane production is
measured. Measuring methane production from live animals is laborious and costly. An
alternative method to estimate methane production from ruminants is by an in vitro technique
(Cone et al., 1996). One main disadvantage of the in vitro technique is that it does not take
into account the dynamic of rumen fermentation and digestion kinetics. On the other hand,
interactive effects among feeds can influence the stoichiometry of rumen fermentation, which
could modify the methane production per unit of dry matter. Recently, Ramin and Huhtanen
(2012) developed an in vitro method to predict ruminant methane production in vivo by the
use of modelling approaches. The objective of the current study was to compare predicted
methane production from the in vitro simulation by incubating feeds either individually or as
total mixed diets.
Materials and Methods
Eleven diets were evaluated, with forage-to-concentrate ratios from 90:10 to 45:55. The
forages (n = 9) were tropical grass pasture plants (Brachiaria sp.) and maize silages (Zea
mays). The concentrates (n = 7) were obtained by mixing whole soybeans (SBW) or soybean
meal (SBM) with ground maize, wheat bran, urea and minerals in different proportions
(Table 1). A fully automated gas in vitro system was used as described by Cone et al. (1996)
with recording of gas production (GP) every 12 minutes. The recorded GP was corrected to
normal air pressure (1013.5 h Pa). Feeds were incubated both individually and as mixed diets
of forage and concentrates in varying ratios. Samples of 1 g were weighed directly into 250ml serum bottles and incubated in 60 mL of buffered rumen fluid for 48 hours. The bottles
were placed in water baths at 39°C and gently agitated continuously during the incubations.
Rumen fluid for all three in vitro incubation runs was obtained from the same two ruminally
cannulated Swedish Red cows about two hours after morning feeding. Cows were fed on a
diet containing grass silage (60%) and commercial concentrates (40%). The rumen fluid was
collected into pre-warmed thermos flasks previously flushed with carbon dioxide and
afterwards filtered through four layers of cheese cloth into a buffered mineral solution
(Menke and Steingass, 1988), with the ratio of rumen fluid to buffer of 1:4 (V/V). All 27
samples (forages, concentrates and total mixed diets) were incubated in three in vitro series
(runs) and were randomly distributed within the runs, resulting in three in vitro observations
per sample. In each run, a blank (buffered rumen fluid without a sample) was incubated in
duplicates. Methane production was predicted as described by Ramin and Huhtanen (2012).
Methane production was reported as weighted means for individual feeds and total diet
Proceedings of the 5th Nordic Feed Science Conference
143
Methods and miscellaneous
separately. The statistical comparison was performed by simple linear regression of values
obtained from total diet incubations (Y) on values obtained from the weighted sum of the
individual feeds (X) of respective diet. The following null hypotheses were tested:
H0 : β0 = 0
(1)
H 0 : β1 = 1
(2)
where β0 is the intercept, and β1 is the slope. The methane production estimates obtained by
diet or individual feeds incubations should be considered similar if both of the null
hypotheses are not rejected. The model adjustment was performed by using the MIXED
procedure of SAS (α = 0.05). The random effect of the different runs was included in the
model.
Results and Discussion
The descriptive statistics for methane production are given in Table 2. The mean of predicted
in vivo methane production from the total diets was 30.1 mL/g DM and 30.8 mL/g DM from
the weighted mean of individual feeds, respectively. There was a poor correlation (r = 0.15)
between weighted methane production from individual feeds and total diets (Table 3 and
Figure 1). Both complete diet and individual feed digestibility, providing energy for rumen
fermentation, do not only depend on individual intrinsic feed characteristics. Interactive effects
between dietary components on digestibility, especially on the neutral detergent fibre (NDF)
fraction, have been reported for tropical diets (Detmann et al., 2008). This could be one
reason for the differences in methane production between observed and predicted values from
individual feeds observed in the current study. In this study, the digestibility of the NDF
fraction showed a quadratic response in the complete diets based on Brachiaria sp to the
replacement 0, 50 and 100% of the SBM with SBW in the concentrate (0.86:0.14
forage:concentrate ratio). Maximum digestible NDF was obtained at 50% SBM replacement
with SBW (611.5 g/kg), compared with 0 (539.6 g/kg) and 100% (548.6 g kg) replacement
(data not shown). On the other hand, the predicted weighted mean of methane production for
the highest NDF digestibility (25.7 mL/g DM) was remarkably different from the observed
value for the total diet (32.2 mL/g DM). The responses were completely opposite and are
supported by the poor relationship shown in Table 3 and Figure 1.
Based on these findings, it is not possible to predict methane production of total diets based
on individual ingredients because their sum is not equal to observed values from the in vitro
incubation of the total diet. The associative effect of feeds on diet digestibility was also
reported by Huhtanen (1991).
It is often assumed that energy values of feeds are additive and that there are no interactions
when they are mixed. However, that might not always be true (Huhtanen, 1991). Associative
effects have occurred when the apparent digestibility of a mixture does not equal the sum of
the separately determined digestibilities of its components (Mould, 1988). Feed interactions
could have been a reason for the differences between predicted methane production from
individual feeds and predicted methane production from the total mixed rations.
144
Proceedings of the 5th Nordic Feed Science Conference
Methods and miscellaneous
Table 1 Proportions of ingredients of 7 concentrates used as substrates for the comparison of methane
production from individual feeds and total diets in the in vitro gas production system.
SBW relative to SBM (3 concentrates)a
4 concentratesb
0%
50%
100%
Ground maize
0
0
0
Wheat bran
43
42.5
41.5
0
Soybean meal
50
25
0
13.8
Soybean grain
0
25
50
0
Minerals
6
6
6
2.6
Urea
1
1.5
2.5
2
Total
100
100
100
100
81.6
a
SBW = whole soybeans; SBM = soybean meal.
The feed composition of concentrates are the same (the ratio of different feeds) but differed in their
chemical composition.
b
Table 2 Descriptive statistics for the methane production (mL/g DM) obtained from incubation of total diet and
from weighted information of individual feeds.
Statistic
Total diet
Weighted value
Mean
30.1
30.8
Minimum
21.3
22.9
Maximum
40.4
38.9
Standard deviation
5.64
4.31
a
n
a
32
One total diet had 2 replicates in the in vitro runs.
Table 3 Estimates of linear regression parameters for the methane production (mL/g DM) obtained from
incubation of total diet (Y) and from weighted information of individual feeds (X).
Parameter
Estimate
s.e.
P value
Intercept
22.1
7.28
0.005a
Slope
0.26
0.235
0.003b
r
0.15
-
0.431 c d
a
H0: β0 = 0. b H0: β1 = 1. c H0: ρ = 0. d The correlation has been adjusted for the random variation among
incubation runs.
Conclusions
It is concluded that the weighted sum of individual feeds cannot be used as a proxy for the
estimation of methane production from total mixed diets.
Proceedings of the 5th Nordic Feed Science Conference
145
Methods and miscellaneous
45
Complete diet CH4 production (mL/g DM)
40
35
30
25
20
Y=X
20
25
30
35
Weighted CH4 production (mL/g DM)
40
45
Figure 1 Descriptive relationship between the methane productions obtained from incubation of total diet and
from weighted information of individual feeds (see Table 3 for details of the relationship).
References
Cone, J. W., Van Gelder, A. H., Visscher, G. J. W. & Oudshoom, L., 1996. Influence of
rumen fluid and substrate concentration on fermentation kinetics measured with a fully
automated time related gas production apparatus. Anim. Feed Sci. Technol. 61, 113-128.
Detmann, E., Paulino, M. F. & Valadares Filho, S. C. 2008. Nutritional evaluation of feeds or
diets? A conceptual approach (in Portuguese). In: Proceedings of II International Symposium
of Beef Cattle Production (II Simpósio Internacional de Produção de Gado de Corte and VI
Simpósio de Produção de Gado de Corte – SIMCORTE). Universidade Federal de Viçosa.
Viçosa, MG, Brazil. p.21-52.
Huhtanen, P., 1991. Associative effects of feeds in ruminants. Norw. J. Agric. Sci., Suppl. 5,
37-57.
Johnson, K. A. & Johnson, D. E., 1995. Methane emissions from cattle. J. Anim. Sci. 73,
2483-2492.
Menke, K. H. & Steingass, H.,. 1988. Estimation of the energetic feed value obtained from
chemical analysis and in vitro gas production using rumen fluid. Anim. Res. and Develop.
28:7‐55.
Mould, F. L., 1988. Associative effects of feeds. In: E. R. Ørskov (ed), World Animal
Science B4, Feed Science. Elsevier, Amsterdam, pp 279-292.
Ramin, M. & Huhtanen, P., 2012. Development of an in vitro method for determination of
methane production kinetics using a fully automated in vitro gas system – A modeling
approach. Anim. Feed Sci. Technol. 174, 190-200.
Ramin, M. & Huhtanen, P., 2013. Development of equations for predicting methane
emissions from ruminants. J. Dairy Sci. 96, 2476-2493.
146
Proceedings of the 5th Nordic Feed Science Conference
Methods and miscellaneous
Relationship between chewing index and intake of metabolizable energy in pregnant
ewes
M. Vestergaard Nielsen1*, E. Nadeau2, B. Markussen3, C. Helander2, M. Eknæs4, Å. Randby4
& P. Nørgaard1
1
Faculty of Health and Medical Sciences, University of Copenhagen, Grønnegårdsvej 3, 1870
Frederiksberg, Denmark. [email protected] and [email protected], 2Department of Animal
Environment and Health, Swedish University of Agricultural Sciences, Gråbrödragatan 19,
SE-532 31 Skara, Sweden. [email protected] and [email protected], 3Laboratory of
Applied Statistics, Faculty of Sciences, University of Copenhagen, Universitetsparken 5, 2100
Copenhagen, Denmark. [email protected], 4Department of Animal and Aquacultural
Sciences, Norwegian University of Life Sciences, P.O.Box 5003, N-1432 Ås, Norway
Correspondence: [email protected]
Introduction
Mathematical models have proven to be a powerful tool for improving animal performance
(Tedeschi et al. 2005). A new model for prediction of feed intake under Scandinavian
conditions could help optimizing ewe performance and, thereby, improve sheep productivity.
Nørgaard and Mølbak (2001) have presented a linear model describing net energy (NE)
intake in relation to metabolic BW as a linear function of the total ration chewing index (CI)
for cattle of different types: lactating dairy cows, non-lactating non-pregnant cows, growing
steers and bulls. This model was able to describe feed intake over three very different
physiological stages, hence the model possibly could be used to describe feed intake also of
other types of ruminants, e.g. pregnant ewes. Nørgaard and Mølbak (2001) hypothesized the
biological explanation of the intercept of the model, to be the theoretical maximum energy
intake capacity of the cattle. Nørgaard and Mølbak (2001) also hypothesized that the rate of
decreasing energy intake with increasing CI is dependent on the theoretical maximum energy
intake capacity, and further discovered a direct proportionality between slopes and squared
intercept values from the individual experiments used in their model development. Increasing
theoretical maximum energy intake capacity results in steeper slope. The greater potential an
animal energy intake, the more sensitive that animal is to the filling effect of the feed.
The aim of this study was to investigate whether the linear relationship between
metabolizable energy (ME) intake and the chewing index (CI) of the rations, also was present
for pregnant ewes. A secondary objective was to test for direct proportionality between the
slopes and squared intercepts when the model was developed for pregnant ewes.
Materials and Methods
The investigation was based on intake data from a total of 101 pregnant ewes, distributed on
11 different dietary treatments based on individual weekly means of energy intake over the
last four weeks before parturition. One Norwegian and four Swedish experiments were
included in the study. All ewes were fed ad libitum grass silage, supplemented with
concentrate fed separately or as a total mixed ration. The ewes were adults of different
breeds; the Swedish experiments included Finewool-Dorset crosses, and the Norwegian
experiment included Norwegian White. Ages of the ewes ranged from two to seven years,
with mean body weight (BW) of 99.6 kg (SD = 10.90), and the ewes were pregnant with 1 to
4 lambs. A summated representation of the data is presented in table 1.
Proceedings of the 5th Nordic Feed Science Conference
147
Methods and miscellaneous
Table 1 Summary characteristics1 of the five experiments. The feed characteristics and BW are presented as
mean of the experiment and minimum and maximum values of the individual animals; the concentrate (conc.) is
listed as amount given or range of amounts given.
C
E T No
NDFf
(g/kg
DM)
CPf
(g/kg
DM)
iNDFf
(g/kg
NDF)
MEf
(MJ/kg
DM)
Conc.
(kg/d)
CI corr
(min/ MJ
ME)
BW
(kg)
S
1
541
150
121
11.4
0.2-0.8
27.8
103.0
(446-643)
(97-182)
(77-164)
(10.3-12.1)
(19.7-46.5)
(84.0-123)
553
156
71
11.0
30.5
98.2
(490-608)
(145-167)
(64-75)
(10.3-11.8)
(24.4-44.5)
(80.5-120)
453
202
11.5
24.7
99.6
(397477)
(183-226)
34 (1842)
(19.5-38.2)
(76.0-122)
473
152
35
11.5
23.9
98.2
(383-510)
(145-175)
(25-41)
(10.9-11.6)
(19.6-28.3)
(82.0-131)
506
132
79
11.2
32.5
102.8
(457-573)
(106-149)
(17-164)
(5.20-14.9)
(20.0-47.0)
(72.8-142)
S
S
S
2
3
4
N 5
2
3
3
3
3
6
7
7
8
412
0.8
0.8
(11.2-12.5)
0.5
0.0-0.4
1
C= country, E= experiment, T= treatments, No=Number of animals per treatment. NDFf is content of NDF per
kg DM of the forage, CPf is crude protein content of the forage, iNDFf is iNDF content of the forage per kg
NDF, MEf is metabolizable energy content of the forage. CI corr respond to the CI values calculated in NorFor
and corrected for NDF intake and BW.
The Nørgaard and Mølbak model considered intake of NE using Scandinavian Feed Units
(SFU) with 7.89 MJ NE/SFU (Nørgaard and Mølbak 2001). Instead of predicting NE intake
for ewes, this study utilized the more easily obtainable ME intake of the ewes. The ME
intake was estimated based on dry matter intake and ME content of the feeds. The ME
content of forages in the Swedish experiments was calculated from in vitro organic matter
digestibility (VOS; Lindgren 1979, 1983, 1988). For concentrates in the Swedish
experiments, the ME was calculated according to Axelsson (1941). In the Norwegian
experiment in vivo digestibility trials were performed and ME was calculated for both
concentrate and forage according to Van Es (1978). The CI was expressed per MJ of ME
intake. The CI of the forages was estimated according to NorFor using neutral detergent
fiber (NDF; g/kg dry matter), indigestible NDF (iNDF, g/kg NDF) and theoretical particle
length (mm; Nørgaard et al., 2011). For some experiments the iNDF content was not
available, and was therefore estimated from the NDF content and in vitro organic matter
digestibility based on the work by Eriksson (2010) and Åkerlind et al. (2011) using iNDF
(g/kg DM) = (-5.7462*VOS(%)+537.23)/(forage NDF (g/kg DM)*1000). The CI were
corrected for BW and for the deviations of forage intake from 0.7% of the BW, which is the
norm in NorFor (Nørgaard et al., 2011). All concentrates were given a CI value of 4 min/kg
DM as described by Nørgaard et al. (2011). The data was standardized as weekly means of
intake for each ewe, the corrected CI values were also calculated as weekly means for each
ewe. Nørgaard and Mølbak (2001) described the NE intake in relation to metabolic BW due
to the different types of cattle and type of production. However, ME intake estimates were
148
Proceedings of the 5th Nordic Feed Science Conference
Methods and miscellaneous
not corrected for metabolic BW in the present model, because only pregnant ewes of
relatively similar size and production were considered.
The data was analysed by linear mixed effects modelling in R (ver. 2.15.2) with random
intercept and slope dependent on experiment (Equation 1). Wald test was used to analyse the
effect of model variables (Sørensen and Ekstrøm 2010), using an assumption of normal
distribution to test a hypothesis about a single parameter, i.e. that intercept is different from
zero. The random effect on slope and intercept was tested by a likelihood ratio test.
MEIi(,j) = interceptj + bj*CIi(,j) + ε i(,j)
(Equation 1)
Where MEIi(,j) is the metabolizable energy intake (MJ ME/day) of the individual ewe in week
i within experiment j, interceptj is the intercept (MJ ME/day) of experiment j, bj is the slope
((MJ ME/day)2/(min/day)) of experiment j, CIi(,j) is the chewing index (min/MJ ME) from the
individual ewe in week i within experiment j, and ε i,(j) is the error of the regression in week i
within experiment j. Note that the unit for the slope (MJ ME/day)2/(min/day)) is to achieve
unit balance of Equation 1. Linearity of the model was analysed by plots of standardized
residuals versus fitted values and q-q plot, which compares the standardized residuals
quantiles to those of the normal distribution (Sørensen and Ekstrøm 2010). To assess the fit
of the model, R2 was calculated, as the squared correlation between observed and predicted
values based on the intercept and slopes including both fixed and random effect. As described
by Nørgaard and Mølbak (2001), the proportionality between the slopes and squared
intercepts values from each of the four weeks before parturition within the experiments was
analyzed by linear modeling (Equation 2). Pearson’s product-moment correlation between
slopes from each experiment and squared intercepts from each experiment was calculated and
was tested for significance by the Wald test (Sørensen and Ekstrøm 2010), the hypothesis
being that the intercept, q (see Equation 2), is not significantly different from zero.
b i(,j) = q+k*intercept i(,j) 2 + ε i(,j)
(Equation 2)
Where b i(,j) is the slope of week i within experiment j, intercept i(,j) 2 is the squared intercept
of week i within experiment j, and ε i(,j) is the error of the regression in week i within
experiment j.
Results and Discussion
A visualization of ME intake as a function of the CI of the rations showed that the data had
large variation, but a pattern could be found suggesting that a linear relationship exists
(Figure 1).
The likelihood ratio test showed significant random effect of both week and experiment on
both slope and intercept values of the model for pregnant ewes, as shown in Table 2. The
intercepts differed between experiments, and for all experiments the trend was for the
intercepts to increase closer to parturition. The slopes did not show such a trend.
The standardized residual plot for Equation 1 showed no pattern of distribution and the q-q
plot showed a linear relationship between quantiles of standard normal residuals versus
standardized residuals, suggesting that the linear relationship between MEI and CI existed for
the pregnant ewes. The R2 for the model was 0.51 using fixed and random effects, suggesting
that the model suffers from the large variation in the data and will have to be improved to
predict satisfactory.
Proceedings of the 5th Nordic Feed Science Conference
149
Methods and miscellaneous
Figure 1 Treatment means with linear regression lines for metabolizable energy intake (MJ/day) of pregnant
ewes in each of the last four weeks pre partum as a function of the rations’ chewing index (min/MJ).
Table 2 Intercept and slope of the linear regression of metabolizable energy intake as a function of the dietary
chewing index, MEI=intercept+b*CIcorr, expressed with fixed effects, standard error and significance for
pregnant ewes.
Parameter
Estimate
SE
P-value
Intercept, MJ ME /day
43.40
3.99
<0.001
Slope (b), ((MJ ME/day)2/(min/day))
-0.517
0.120
<0.001
With regards to direct proportionality between the slope and the squared intercept in Equation
1, Figure 2 displays a scatterplot of the corresponding estimates from the five experiments. A
linear regression was made between the slopes and the squared intercepts. Using Wald test
for evaluation, the resulting linear regression, had an intercept not significantly different from
zero (P=0.17) indicating that equation 2 is true. This suggests that the slope can be described
as a function of the squared intercepts.
Nørgaard and Mølbak (2001) suggested that the intercept values from the individual
experiments represent a corresponding theoretical capacity for energy intake, as it occurs
when no dietary constraint on intake exists, for CI = 0. The theoretical capacity for energy
intake of the ewes is considered to be exclusively related to animal characteristics (Elsen,
1988; Nørgaard and Mølbak, 2001). The described pattern of increasing intercepts of the
model closer to parturition shows increased theoretical maximum intake capacity closer to
parturition. Using animal characteristics to predict intercepts, and using forage characteristics
to predict the slopes might improve predictability, and thereby improve the practical
application of this linear relationship. Since direct proportionality is present between slopes
and squared intercepts, any improvement in the prediction of the intercepts, would also
impact the predicted slopes of the model, and might improve the model further.
150
Proceedings of the 5th Nordic Feed Science Conference
Methods and miscellaneous
Figure 2 Plot of slopes and squared intercepts found with Equation 1 for pregnant ewes the last four weeks
before parturition, the line represents the linear regression to illustrate linearity and intercept, which was not
significantly different from zero (P=0.17).
Conclusions
This study supports the hypothesis that ME intake of pregnant ewes during the last four
weeks before parturition decreases linearly at increasing dietary chewing index. The slope
values appear to be proportional with the squared intercepts values. The intercept values are
considered to reflect the metabolic capacity of the ME intake by pregnant ewes. Thus, the
model proposed by Nørgaard and Mølbak (2001) could be relevant also for pregnant sheep.
However, the potential prediction power of the model was unsatisfactory and the model needs
further improvements.
References
Axelsson, J., 1941. Der Gehalt des Futters an umsetzbare Energie. Z. Züchtungsk. 16, 335–
347.
Elsen, J. M., Wallach, D. & Charpenteau. J. L., 1988. The calculation of herbage intake of
grazing sheep: A detailed comparison between models. Agric. Syst. 26, 123-160.
Eriksson, T., 2010. In vitro methods for indigestible Neutral Detergent Fiber (iNDF): Proc.1st
Nordic Feed Science Conference, Uppsala, Sweden, 22-23 June, 2010.
Lindgren, E., 1979. The nutritional value of roughages determined in vivo and by laboratory
methods. Report 45. Dep. Anim. Nutr. Managem., Swed. Univ. Agric. Sci., Sweden (in
Swedish with English summary), 66 p.
Lindgren, E., 1983. Nykalibrering av VOS-metoden för bestämning av energivärde hos
vallfoder. Working paper. Dep. Anim. Nutr. Managem., Swed. Univ. Agric. Sci., Sweden (in
Swedish), 4 p.
Lindgren, E., 1988. Fodrets energivärde. Course paper Feed Science HNU 3. Dep. Anim.
Nutr. Managem., Swed. Univ. Agric. Sci., Uppsala, Sweden (in Swedish). 49 p.
Proceedings of the 5th Nordic Feed Science Conference
151
Methods and miscellaneous
Nørgaard, P. &. Mølbak, L., 2001. Relation between dietary chewing index value and net
energy intake in cattle fed concentrates restrictively and forage ad libitum. EAAP
Publication. 103, 67-70.
Nørgaard, P., Nadeau, E., Randby, Å. & Volden, H., 2011. 11. Chewing index system for
predicting physical structure of the diet. EAAP publications No. 130 (eds. H. Volden).
Wageningen Academic Publishers, The Netherlands, 2011, p. 127-132.
Sørensen, H. & Ekstrom, C.T., 2010. Introduction to Statistical Data analysis for the Life
Sciences. CRC Press.
Tedeschi, L. O., Fox, D. G., Sainz, R. D., Barioni, L. G., d. Medeiros, S. R. & Boin, C, 2005.
Mathematical models in ruminant nutrition. Scientia Agric. 62, 76-91.
Van Es, A. J. H., 1978. Feed evaluation for ruminants. I. the systems in use from May 1977
onwards in the Netherlands. Livest. Prod. Sci. 5, 331-345.
Åkerlind, M., Weisbjerg, M., Eriksson, T., Tøgersen, R., Udén, P., Ólafsson, B.L., Harstad,
O.M & Volden, H., 2011, 5. Feed analyses and digestion methods. EAAP publications No.
130 (eds H. Volden). Wageningen Academic Publishers, The Netherlands, 2011, p. 41-54.
152
Proceedings of the 5th Nordic Feed Science Conference
Methods and miscellaneous
In-line oxygen monitoring in lab-scale silos
T.M. Pauly
Swedish University of Agricultural Sciences (SLU), Department of Animal Nutrition &
Management, Kungsängen Research Centre, 753 23 Uppsala, Sweden.
Correspondence: [email protected]
Introduction
The principle of lab-scale ensiling experiments is to apply one or
several specific treatments to homogenized forage or a group of
silos while all important environmental factors, which might
affect fermentation quality, are kept as equal as possible among
silos. This way, a specific treatment effect can be studied with
little interference from confounding factors. However, in many
experiments, control over the ingress of air during the storage of
silos (or wrapped bales) is either ignored or not reported despite
the fact that the deteriorating influence of oxygen on fermentation
quality of silage is well described (Honig, 1994; Ruxton et al.,
1975; Tabacco et al., 2009; Woolford, 1990). This leads to
unnecessary large response variations and clouds potential
differences among treatments.
At our research facility, we routinely use water-filled plastic
siphons, which are attached to the lids of our mini-silos (Figure 1)
and allow fermentation gases to escape (as bubbles) from the silos
without the ingress of air. In addition, the level of the two water
columns indicates if a silo was badly sealed (water columns in
level) or if a silo was tight (water columns on different levels indicating an over- or underpressure in the silo). This has helped us to
remedy badly sealed silos, increased accuracy of our studies and
allowed us occasionally to reduce the number of silo replicates.
Figure 1 Glass jar silo
with a water-filled
siphon and 2 stoppers
(top+bottom, see
arrows), which were
used to inject air into
the silo.
Our goal is not to exclude any oxygen from our mini-silos but to
create the same degree of air leakage in all silos and, if possible, to quantify the amount of
oxygen leaking into our silos. For this purpose, we recently bought oxygen sensor spots
(from: www.pyro-science.com), which easily fit in our mini-silos and which can record
oxygen concentrations over a wide range in both gas and liquid phases. Below, it is shown
how this equipment is used in a small pilot study.
Materials and Methods
A single glass jar was filled with freshly mown loan grass (23% DM) to get an indication on
how long time it took to deplete the oxygen concentration in the silo after sealing. At a later
occasion (Oct. 24th), three glass jar silos (1.8 L volume) were filled with precision-chopped
maize forage (41% DM) and injected them with either 4.25, 8.5 or 17.0 mL of air (equivalent
to 0.2%, 0.5% and 0.9% of total silo volume). Air was injected by hand with a syringe and a
thin hypodermic needle (0.7 x 30 mm) 3 times per week (Mon, Wed, Fri). Hypodermic
needles were inserted through either the bottom or the top stopper of the silo and they were
not removed during the study. Needle ends were plugged with silicone stoppers (except
during injections) to avoid any air leakage (Figure 2). When air was injected through the top
Proceedings of the 5th Nordic Feed Science Conference
153
Methods and miscellaneous
stopper in the silo lid, the siphon was plugged with a
stopper (to avoid loss of air) and the bottom stopper was
removed from the needle (to release overpressure). Our
intention was to study the decrease of oxygen concentration in the silo after air injections to get an indication on
how fast oxygen might be consumed in silage. Air was
assumed to contain 20.5% oxygen.
glass jar
silo
stoppers
A vital part of the oxygen monitoring set is the oxygen
sensor spot (Ø = 5 mm), which contains a green, oxygensensitive dye. Sensor spots were glued on the inside of
each silo right below the lid, i.e. in the silo’s head space.
On the opposite side of the glass wall, a fibre cable was
attached with a special adapter. This fibre cable ended in
a device (FireStingO2) that emitted red light of a specific
wave length and detected the IR light reflected from the
sensor spot. A change in the silo’s oxygen concentration
hypodermic
needle
Figure 2 Glass jar silo (approx. 1.8
L volume) with hypodermic needle
inserted through the bottom
stopper. The needle was sealed by
another stopper.
changed the reflection properties of the dye, which was
then translated by the FireStingO2 device into a specific
oxygen concentration. Before the onset of each
recording, the whole setup was calibrated with <0.01%
O2 (in a tightly sealed silo) and air (20.5% O2). A
temperature probe corrected automatically for temperature changes during measurements. Measurement
intervals and recorded values were saved and displayed
graphically on a connected computer.
transparent
silo wall
Results and Discussion
Decrease of oxygen concentration in a freshly sealed
mini-silo with unwilted loan grass
Oxygen (%) in silo
In the unwilted grass, forage oxygen concentration
fell from 20.5% to below 0.1% within 47 min.
We assume that this decrease was mainly
caused by plant respiration. The decrease was
rather linear at approximately 4.4%-units for
every 10 min. (Figure 4). Sprague (1974)
reported that when he filled fresh alfalfa (24%
DM) in mini-PVC-silos, about 0.5% O2
remained 30 min. after sealing, which was a
slightly faster O2 decrease than in our silo.
Figure 3 Detection principle of an
oxygen sensor spot (green) containing
an oxygen sensitive dye.
25
20
15
10
5
0
0
10
20
30
40
Time (min)
50
60
Figure 4 Decrease of oxygen concentration in
freshly sealed grass forage (23% DM).
154
Proceedings of the 5th Nordic Feed Science Conference
Methods and miscellaneous
Decrease of oxygen concentration in 3 air-injected silos
We started to inject air through the bottom stopper but were unable to get any oxygen
response from the sensor spots, not even after the injection of 17 mL of air (3 attempts and
waiting for 30 min after injection). Therefore, injections through the top stoppers into the silo
head space were done instead. By doing so, almost instantaneous and sharp oxygen peaks,
proportional to the quantity of injected air, were seen (Figure 5).
After the injection peak, oxygen concentrations dropped at an exponential rate approaching
zero (Figure 5). This change was different from the almost linear oxygen decrease after
sealing (Figure 4) and was evident after each of the 3 air injection levels. The reasons for this
might be that there are two underlying processes influencing oxygen concentration in the
head space of the silo: a) the dispersion of oxygen in the head space and penetration into
silage pores (rel. fast O2 decrease by dilution) and b) the consumption of oxygen by microorganisms (rel. slow O2 decrease). The combination of these two processes might be
responsible for the oxygen curves shown in Figure 5.
3.5
Oxygen concentration (%)
3.0
17.0 mL air
17.0 mL air
2.5
2.0
1.5
1.0
8.5 mL air
0.5
4.25 mL
0.0 air
-5
0
5
10
15
20
25
Time (min)
30
35
40
45
Figure 5 Decrease of oxygen concentration in the head space of maize silage (41% DM) 18 days after ensilage.
3 different quantities of air were injected.
Another explanation might be that oxygen-consuming microorganisms and oxygen are not
evenly distributed within a silo. In case oxygen-consuming microorganisms have depleted
oxygen at one spot, oxygen has to diffuse from surrounding areas, which takes time,
particularly when the O2 concentration gets low (i.e. low partial pressure difference).
Proceedings of the 5th Nordic Feed Science Conference
155
Methods and miscellaneous
Conclusions
• Oxygen disappears quickly (<1 h) from a freshly sealed silo with unwilted forage due to
plant respiration.
• The reason why we didn’t get any oxygen response from the sensor spots after injections
through the bottom stopper is most likely that the oxygen was completely absorbed and/or
metabolised on its way from the place of entry (bottom stopper) to the sensor spot (head
space below silo lid; distance bottom stopper to sensor spot approx. 160 mm).
• Only when the oxygen sensor was placed in the head space above the silage (and not in the
silage mass), a clear response of an oxygen leakage can be observed.
• Air injections into the silage head space lead to oxygen peaks proportional to the quantity
of injected air. Oxygen levels decrease at an exponential rate as oxygen concentration
approaches zero.
References
Honig, H., 1994. Amount and effects of gas exchange in silage stacks. Proc. of the 15th Gen.
Meet. Europ. Grassl. Fed., 6-9 June 1994, Wageningen, The Netherlands. p.143-145.
Ruxton, I.B., Clark, B.J. & McDonald, P., 1975. A review of the effects of oxygen on silage.
J. Brit. Grassl. Soc., 30, 23-30.
Sprague, M.A., 1974. Oxygen disappearance of alfalfa silage (Medicago sativa L.). Proc. of
the 12th Internat. Grassl. Congr., Moskva, p.216-225.
Tabacco, E., Piano, S., Cavallarin, L., Bernardes, T.F. & Borreani, G., 2009. Clostridia spore
formation during aerobic deterioration of maize and sorghum silages as influenced by Lactobacillus buchneri and Lactobacillus plantarum inoculants. J.Appl. Microbiol., 107, 1632–
1641.
Woolford, M.K., 1990. A review: The detrimental effects of air on silage. J.Appl. Bacteriol.,
68, 101-116.
156
Proceedings of the 5th Nordic Feed Science Conference
Methods and miscellaneous
Effects of heat treatment on protein feeds evaluated in vitro utilizing the method utilizable
protein
M. Vaga, M. Hetta & P. Huhtanen
Dept.of Agricultural Research for Northern Sweden, Swedish University of Agricultural Sciences, SE901 83 Umeå, Sweden.
Correspondence: [email protected]
Introduction
There is an increasing interest in using domestic feeds in diets for dairy cattle in Sweden. Due
to increasing energy prices as well as ethical considerations and growing demand for certified
organic feeds it is not rational anymore to depend on imported protein supplements for dairy
cattle. Seeds of field beans (Vicia faba), and lupines (L. angustifolius ) have high crude
protein (CP) and starch content (NRC, 2001) and therefore have potential to replace imported
soybean meal (May et al. 1993; Tufarelli et a. 2012).
Meeting dietary protein requirements of high producing dairy cows rely largely on the
proportion of ruminally undegraded protein which also has high total digestibility. Heat
treating protein feeds have been suggested to decrease protein degradability and increase
nitrogen efficiency in dairy production (Broderic & Graig, 1980; Cros et al. 1991; Canbolat et
al. 2005). One of the challenges has been to evaluate the effects of heat treatments with
relevant laboratory methods as production trials are expensive. Several new methods for
evaluation of forage protein for ruminants have been presented, e.g. Karlsson et al. (2009).
The technique described by Karlsson et al. (2009) originates from the method developed by
Raab et al. (1983), which utilises the production of ammonia nitrogen (NH3-N) for the
estimation of protein degradation in vitro. The method has also been modified into another
promising method for estimation of protein supply from forages by using utilisable crude
protein (uCP) (Edmunds et al., 2012). It quantifies simultaneously the availability of protein
as the sum of the microbial de novo synthesis of protein in combination with estimations of
the degradation of feed protein resulting in a total estimate of supply of protein from the
rumen to the small intestine. Using in vitro methods for protein evaluation remove the
problems like escape of soluble non-ammonia-N and particle losses, which occur in in situ
methods (Ørskov and McDonald, 1979). The aim of this experiment was to evaluate the
effects of heat treatment on protein feeds utilising the in vitro method utilisable protein.
Materials and Methods
Feeds and Experimental design
The experiment had 2x2x4 factorial design plus one control, including two protein feeds, two
heat treatment methods and three temperature levels + untreated and one control diet,
resulting in a total of 17 treatments. The two protein feeds used in this experiment were
grains of field beans and lupines. Each protein feed was mixed together with standard silage
and barley in isonitrogenous total mixed rations (TMR) in order to simulate real diets and
take into account effects on microbial protein synthesis which provides about two thirds of
the of the protein supply to the ruminant. The TMR diets consisted of 500 mg of standard
grass silage, barley and protein feed, combined to give a CP concentration of 180 g/kg DM
(Table 1). The control diet consisted of only standard silage and barley (1:1, fresh weight)
with total CP concentration of 132 g/kg DM. Each diet was randomly distributed within and
between in vitro series (run) and replicated twice within two of total eight runs, resulting in
Proceedings of the 5th Nordic Feed Science Conference
157
Methods and miscellaneous
four observations per sample. In each run, a blank (buffered rumen fluid without a sample)
and the control diet were incubated in duplicates.
Sample preparation and heat treatments
The feeds were dried at 60°C in force air oven for 48h and milled through a 1mm screen
(cutter mill; SM300, Retsch GmbH, HAAN, Germany). Thereafter the samples of protein
feeds, silage and barley were analysed for dry matter (DM) (105oC for 16 h), ash (AOAC
1984) and CP (AOAC 1984). The chemical composition of the feeds and diets are presented
in Table 1.
Table 1 Chemical composition of experimental feeds and diets
Protein feeds
Diets
Silage
Barley
FB
LP
FBD
LPD
Control
DM, g/kg
939
866
905
882
914
907
903
OM, g/kg DM
941
969
960
959
956
956
960
CP, g/kg DM
143
120
299
348
180
180
132
FB – field beans; LP – lupines; FBD – field beans diet (field beans meal+silage+barley); LPD – lupines
diet (lupines meal+silage+barley); Control - silage+barley diet.
The two protein feeds were heat-treated for 60 minutes at 120, 140 and 160°C in a ventilated
oven (ULE 400, Memmert, Germany) with dry air or at 105, 120 and 135°C in an autoclave
(PACS 50, Getinge Biofoe, Sweden) with steam pressure. The milled feed samples were
placed on aluminium trays and evenly spread (2 cm deep layer). After the heat treatments, the
samples were allowed to cool in open air for 24 h to regain original moisture content (avg.
92% DM) before packing and storing. A sample of each unheated feed was retained for use in
the control treatments.
In vitro procedures
Rumen fluid was collected about two hours after morning feeding from two dry and
ruminally cannulated Swedish red cows fed on a diet containing grass silage (600 g/kg DM)
and commercial concentrates (400 g/kg DM). The rumen fluid was filtered through four
layers of cheese cloth into a buffered mineral solution (Menke and Steingass, 1988), with a
1:4 volume ratio of rumen fluid to buffer. About 1 g of each TMR diet was incubated in 60
ml of buffered rumen fluid in 250-ml serum bottles (Schott, Mainz, Germany). The bottles
were placed in water baths at 39°C for 48 h and continuously agitated. The in vitro gas
production (GP) was recorded as described by Hetta et al. (2003). The in vitro procedures
were performed with a fully automated system (Cone et al., 1996), recording GP every 12
minutes. Simultaneously with gas recordings, determinations of NH3-N concentrations in the
liquid phase were made at 8 h, 24 h and 48 h. The concentration of NH3-N of the buffered
rumen fluid in the incubation bottles was determined by sampling 0.4 ml of fluid with plastic
syringes as described by Karlsson et al. (2009). The fluid samples were transferred into
Eppendorf tubes kept on ice and thereafter, 0.016 ml of 96% H2SO4 was added for
preservation. The samples were stored frozen (-20°C) until analysis. The sample tubes were
thawed, centrifuged (12500 x g, 10 min) and 0.1 ml of supernatant was transferred to test
tubes and diluted 1:20 with distilled water. The concentrations of NH3-N was analysed using
a continuous flow analyser (AutoAnalyzer 3 HR, SEAL Analytical Ltd).
158
Proceedings of the 5th Nordic Feed Science Conference
Methods and miscellaneous
Data handling and statistical analysis
For each treatment, we calculated the concentrations of uCP as described by Edmunds et al.
(2012) (eqn. 1) as:
uCP (g/kg DM) =
NH3Nblank+Nsample−NH3Nsample
x
weight (mg DM)
6.25 x 1000 (eqn. 1),
where NH3Nblank is the average amount (g) NH3N in the two blanks at the time of interest,
Nsample is the amount (g) of N in the sample at the start of the incubation and NH3Nsample is
amount (g) of NH3N in the incubation bottles at the time of interest for the treatment
evaluated.
The effect of heat treatments on uCP and NH3-N were analysed using the GLM procedure in
SAS (SAS Institute 2008). The model included feed, run, treatment temperature and method.
Treatment effects were considered to be significant at P≤0.05.
Results and discussion
The use of heat treatments effectively increased the uCP concentrations for both the field
bean and lupine diets. The lowest temperature for both the oven- and autoclave-treatments
actually reduced the uCP concentrations for the field bean diets after 8 h. This reduction was
expected as it is generally known that modest heat treatments can improve protein
degradability (Fennema, 1976). Increasing the temperatures for oven-treated field beans
caused linear increase of uCP based on the 24 h incubations (Table 2). However for lupines,
only the medium and high temperatures had positive effects, but low temperature treatments
instead decreased the concentration of uCP.
The lupines were less affected by the heat treatments than the field beans. The high
temperature oven-treatment increased uCP levels from 120 to 166 g/kg DM for the field bean
diets and from 167 to 187 for the lupine diets at 24 h. For the autoclaved feeds the increases
were from 120 to 173 g/kg DM for field beans and from 167 to 178 g/kg DM for lupines.
Similar effects were reported by Martinussen et al. (2013) who found that Lupines required
higher temperatures than field beans to achieve the same reducing effect on soluble protein.
Despite the different temperatures, both oven and autoclave treatments seemed to have
similar effect on uCP and NH3-N concentrations in this study.
Changes in concentrations of NH3-N 0, 8 and 24 h incubations are shown in Figure 1. The
autoclave treatments seemed to decrease NH3-N concentrations already after 8 h while for
oven treated feeds, the effect of heat treatment was significant after 24 h of incubation. These
results are supported by Tagari et al. (1986) who noted that autoclave treatment of feeds
increased the NH3-N concentrations at lower temperatures compared to dry heat treatments.
Heat treatment with both oven and autoclave significantly (P < .001) decreased the NH3-N
concentrations as shown in Table 2.
Proceedings of the 5th Nordic Feed Science Conference
159
Methods and miscellaneous
Table 2 uCP and NH3-N concentrations of untreated and treated field beans and lupines diets after 24 hour
incubations
Untreated
Temperature, °C
Oven treated
SE
P
120°C 140°C 160°C
Autoclave treated
SE
P
105°C 120°C 135°C
uCP 24h (g/kg DM)
Field beans
121
141
158
167
18.7 ***
168
158
173
18.7 ***
Lupines
167
153
179
187
17.2 ***
171
190
179
18.8 ***
Control
166
NH3-N 24h (mg/L)
Field beans
451
405
357
340
42.6 ***
317
363
320
42.6 ***
Lupines
341
373
313
290
40.6 ***
331
323
309
47.7 ***
Control
226
Blank
309
*** P<0.001
Figure 1 Effect of heat treatments of field beans and lupines on NH3-N concentrations in vitro.
There were higher correlations (r = -0.97 and r = -0.80) between temperature and NH3-N
concentrations for oven-treated and autoclave-treated lupines than for field beans (r = -0.66
and r = -0.70).
The concentrations of NH3-N (mg/L) had high variance among and within runs at 8 h
(SE=24) and 24 h (SE=42), but very small variance at 0 h (SE=1). Small variance at the start
160
Proceedings of the 5th Nordic Feed Science Conference
Methods and miscellaneous
of the incubations when no interactions had occurred yet, suggest a potential for good
precision of the analytical method.
Conclusions
This study shows that heat treatment was effective in increasing the concentrations of uCP in
both field beans and lupines, but lupines needed higher temperatures to achieve the same
effect as in field beans. The new analytical method provided by Edmunds et al. (2012) for
estimating uCP seems to have good potential for evaluating protein quality in vitro. However
more studies need to be carried out to evaluate the accuracy of this method. Especially, we
need to evaluate the effects of heat treatments in production experiments with dairy cows in
order to evaluate the relevance of the in vitro methods for potential use in feed evaluation
systems.
References
AOAC, 1984. Official methods of analysis of AOAC. Washington DC, USA: Association of
Official Analytical Chemists.
Broderick, G.A. & Craig W.M., 1980. Effect of heat treatment on ruminal degradation and
escape, and intestinal digestibility of cottonseed meal protein. J. Nutr. 110, 2381.
Canbolat, O., Kamalak, A., Efe, E., Sahin, M. & Ozkan, C.O., 2005. Effect of heat treatment
on in situ rumen degradability and in vitro gas production of full-fat soyabeans and soyabean
meal. South African J. Anim. Sci. 3, 186-194.
Cone, J.W., van Gelder, A.H., Visscher, G.J.W. & Oudshoorn, L., 1996. Influence of rumen
fluid and substrate concentration on fermentation kinetics measured with a fully automated
time related gas production apparatus. Anim. Feed Sci. Technol. 61, 113–128.
Cros, P., Vernay, M. & Moncoulon, R., 1991. In situ evaluation of the ruminal and intestinal
degradability of extruded whole horse beans. Reproduction Nutrition and Development, 31,
249–255.
Edmunds, B., Südekum, K.-H., Spiekers, H., Schuster, M. & Schwarz, F.J., 2012. Estimating
utilisable crude protein at the duodenum, a precursor to metabolisable protein for ruminants,
from forages using a modified gas test. Anim. Feed Sci. Technol. 175, 106-113.
Fennema, O.R., 1976. Principles of Food Science. Part I. Food Chemistry. Marcel Dekker,
Inc., New York, NY.
Hetta, M., Cone, J.W., Gustavsson, A.-M. & Martinsson, K., 2003. The effect of additives in
silages of pure timothy and timothy mixed with red clover on chemical composition and in
vitro rumen fermentation characteristics. Grass Forage Sci. 58, 249–257.
Karlsson, L., Hetta, M., Udén, P. & Martinsson, K., 2009. New methodology for estimating
rumen protein degradation using the in vitro gas production technique. Anim. Feed Sci.
Technol. 153, 193-202.
Martinussen, H., Jørgensen, K.F., Strudsholm, F. & Weisbjerg, M.R., 2013. Heat treatment
increases the protein value in Faba beans and Lupins. Proceedings of the 4th Nordic Feed
Science Conference, 287, 34-38.
Proceedings of the 5th Nordic Feed Science Conference
161
Methods and miscellaneous
May, M.G., Otterby, D.E., Linn, J.G., Hansen, W.P., Johnson, D.G. & Putnam, D.H., 1993.
Lupins (Lupinus albus) as a protein supplement for lactating Holstein dairy cows. J. Dairy
Sci. 76, 2682–2691.
Menke, K.H. & Steingass, H., 1988. Estimation of the energetic feed value obtained from
chemical analysis and in vitro gas production using rumen fluid. Animal Research and
Development, 28, 7-25.
NRC, 2001. Nutrient Requirements of Dairy Cattle, 7th ed. National Academy Press,
Washington, DC, USA.
Ørskov, E.R. & McDonald, I., 1979. The estimation of protein degradability in the rumen
from incubation measurements weighted according to the rate of passage. J. Agr. Sci. 92,
499–503.
Raab, L., Cafantaris, B., Jilg, T. & Menke, K.H.,1983. Rumen protein degradation and
biosynthesis. 1. A new method for determination of protein degradation in rumen fluid in
vitro. Brit. J. Nutr. 50, 569–582.
SAS, 2002. User’s Guide: Statistics. SAS Institute, Inc., Cary, NC, USA.
Tagari, H., Pena, F. & Satter, L.D., 1986. Protein degradation by rumen microbes of heattreated whole cottonseed. J. Anim. Sci. 62, 1732.
Tufarelli, V., Khan, R.U. & Laudadio, V., 2012. Evaluating the suitability of field beans as a
substitute for soybean meal in early-lactating dairy cow: production and metabolic responses.
J. Anim. Sci. 83(2), 136-40.
162
Proceedings of the 5th Nordic Feed Science Conference
Methods and miscellaneous
Grass silage extract, feed component suitable for pigs – prospects for on farm
biorefinery
A. Seppälä1, S. Kyntäjä1, L. Blasco1, M. Siika-Aho2, S. Hautala3, O. Byman3, H. Ilvesniemi3,
H. Ojamo4, M. Rinne1, M. Harju5
1
MTT Agrifood Research Finland, 31600 Jokioinen, Finland; 2VTT Technical Research
Centre of Finland, Espoo, Finland; 3Finnish Forrest Research Institute, Vantaa, Finland;
4
Aalto University, Espoo, Finland; 5Valio Ltd, Helsinki, Finland
Correspondence: [email protected]
Introduction
There are approximately 470 000 hectares of grass in Finland that are not fully utilized
(Seppälä et al., 2014). The potential of surplus grass or grass silage as raw material for green
biorefineries has been evaluated in several research projects during last decades (e.g. Grass
2004; Kamm and Kamm 2004; Mandl 2010; Sieker et al., 2011), and demo or pilot plants
have been built in Austria, Germany, Netherlands and Switzerland (Keijsers and Mandl,
2010). Due to different national interests, the explored process chain and the focused end
products have varied including ethanol (Sieker et al., 2011), insulation boards (Grass, 2004),
purified lactic acid and amino acids (Ecker et al., 2012) or biogas and insulation (O’Keeffe et
al., 2011).
During ensiling, grass material undergoes proteolysis mainly caused by plant enzymes
(McDonanld et al., 1991). Because of the proteolysis during the preservation, fresh grass is
often regarded as a better raw material for protein separation than silage (Kammes et al.,
2011). However under Nordic conditions, the short availability of fresh grass is a severe
drawback. The extent of proteolysis during ensiling can be reduced by rapid drop of pH
(McDonald et al., 1991) enhanced by the use of efficient additives. The solubilisation of
protein during ensiling might actually alleviate separation of amino acids. Choi et al., (2003)
measured in primary growth grass silage 427 g free amino acids, 188 g peptides and 53 g
soluble true protein /kg total N meaning that 667 g/kg total N should be relatively easily
separated from the fibrous part and being useful for monogastric animals.
Grass silage (Presto et al., 2013) or silage effluent (Patterson and Kilpatrick, 1991) can be fed
to pigs as such although with some restrictions. However, the authors of this paper have no
knowledge of scientific articles reporting on farm processing of grass silage to produce feeds
for pigs, although the modern production systems (including liquid feeding and automated
systems for dealing with silage, see e.g. www.Pellon.com) would enable automatic
processing of silage before feeding. This paper focuses on the potential for on farm
processing of silage for pigs. The argumentation for more research in this field is based on
results of two preliminary experiments and literature.
Material and methods
Exp1. In the first experiment, an unfertilized timothy-red clover sward was harvested in a
leafy growth stage (primary growth) with or without prewilting and ensiled using a formic
acid based additive (AIV 2 Plus, Kemira Ltd., Helsinki, Finland), The amount of additive
used was 0, 6, 12 or 18 l/t, but the combination “0 l/t on prewilted” was omitted due to
restricted capacity. Three replicate silos were filled per treatment. After ensiling, the juice
was separated from silages by a pneumatic press (7 bars). Silage quality was analysed using
the ARTTURI® silage analysis (Valio Ltd., Helinki, Finland), where dry matter (DM) was
Proceedings of the 5th Nordic Feed Science Conference
163
Methods and miscellaneous
measured by oven drying, fermentation quality was analysed using titration method and
silage composition using NIR method. The silage press juice was analysed for DM (freeze
drying), ash (500°C, overnight) and nitrogen at MTT laboratory. From two samples
representing the treatments “no prewilting-no additive” and “no prewilting - acid 6 l/t”, three
replicates were combined and amino acid profiles were analyzed (Eurofins, Finland).
Exp2. The second experiment was part of a larger project “Protein from grass” but only the
first part of the work is described here. The project was funded by TEKES and Valio Ltd.
Single batch of farm scale precision chopped silage was used as raw material for the process.
The silage was a second cut of timothy-meadow fescue sward ensiled using a formic acid
based additive (AIV 2 Plus 5 l/t, Kemira Ltd., Helsinki). The silage juice was separated using
three different methods; either by the same pneumatic press as in the Exp1, or by
combination of pressing and pressurized hot water extraction (PHWE, pilot scale, 60°C, 10
kg/min, 90 min), or by hot water extraction (HWE, pilot scale, 55°C, three batches of water,
used 2805 kg water/750 kg silage). After the HWE, the extract was concentrated by
evaporation.
Results
Exp1. Botanical composition of the silage was timothy 0.600, red clover 0.337, meadow
fescue 0.055 and unsown species 0.008. On average, the silages had a crude protein (CP)
concentration of 136 g/kg DM, neutral detergent fibre (NDF) 422 g/kg DM, digestible
organic matter 735 g/kg DM and ash 56.6 g/kg DM. Ammonium N and soluble N
concentrations of the formic acid treated silages were low (Table 1). The press juice
represented 0.30 to 0.54 of the wet silage, equivalent to 0.15 to 0.32 of DM of the silage. The
juice had a DM content of 100-150 g/kg and an ash content of 97–140 g/kg DM. The CP
content in the juice (in which amino acids were determined) was 127 or 182 g/kg DM, having
a lysine concentration of either 5.8 (not prewilted, acid 6 l/t) or 4.4 g/100 g CP (not prewilted,
no additive) respectively.
Exp2. The silage had a DM-content of 241 g/kg, ash 115, crude protein 138, NDF 524 and
digestible organic matter 646 g/kg DM, pH 4.19, lactic acid 69 g/kg DM, ethanol 29 g/kg DM
and ammonium N 69 g/kg total N. Results (mass balances) of silage fractioning by different
methods are presented in Table 2. The concentrated silage extract had 536 g DM/kg, 306 g
ash/kg DM, 247 g CP/kg DM and a lysine content of 3.68 g/100 g CP.
164
Proceedings of the 5th Nordic Feed Science Conference
Methods and miscellaneous
Table 1 Silages used as raw material for juice separation by pressing in Exp1 and results of juice separation.
Prewilting
Level of acid (L)
Not prewilted
Prewilted (PW)
DM 220 g/kg
DM 283 g/kg
0
6
12
18
6
12
18
pH
3.7
3.8
3.9
3.8
3.8
4.1
4.3
NH3 N, g/kg total N
34.3
5.0
7.3
13.3
5.0
14.3
Soluble N, g/kg total N 500
329
310
331
274
Lactic acid, g/kg DM
104.0 71.7
64.3
72.7
Volatile fatty acids,
g/kg DM
33.0
16.3
18.7
Sugars, g/kg DM
45.7
121.7
61.3
SEM
Statistical effects
L
PW
LxPW
0.05
***
***
**
23.3
2.16
***
**
NS
273
320
14.7
***
*
NS
63.3
47.0
49.0
3.80
***
***
NS
16.0
24.3
34.0
29.0
2.67
***
***
NS
96.0
82.3
63.3
60.0
18.40
NS
NS
NS
Silage quality
Extraction rates (juice/silage)
Fresh
0.519 0.542
0.497 0.453
0.347 0.313 0.297 0.0121
***
***
NS
Dry matter
0.254 0.283
0.231 0.221
0.179 0.134 0.136 0.0090
***
***
NS
Crude protein
0.291 0.225
0.195 0.178
0.142 0.118 0.119 0.0069
***
***
NS
Ash
0.526 0.505
0.509 0.454
0.378 0.287 0.292 0.0150
**
***
*
Table 2 Mass balances of silage fractioning by different methods. Values are percentages of the original
amount.
Pressing
Sequential combination of
pressing and PHWE2
HWE3 with mixing
Liquid
Solid
Liquid
Solid
Liquid
Solid
30
70
-
-
-
-
9
91
25
75
23
77
Ash
23
77
60
40
-
-
Protein
17
83
30
70
42
58
Lactic acid
30
70
Phosphorus
38
62
79
21
-
Mass
Dry matter
1
1
2
3
Corrections for loss of volatiles not made; PHWE = pressurized hot water extraction; HWE = hot water
extraction.
Proceedings of the 5th Nordic Feed Science Conference
165
Methods and miscellaneous
Discussion
Towards optimal process to fractionate grass silage
Cell solubles (DM minus NDF) represent nearly half of the DM of high digestibility grass
silage. Results of Exp1 suggest that more than half of cell solubles could be separated by
simple pressing when the leafy silage was not prewilted but ensiled with a formic acid based
additive using 6 l/t. Further, the results of Exp2 suggest that the gain would be improved by
using hot water extraction instead of pressing. Over 80% of organic matter of leafy grass
silage can be solubilised by the combined use of pepsin and cellulases in a laboratory process
(Huhtanen et al., 2006). The fractioning could possibly also be facilitated by the use of
enzymes.
Protein value of silage extract
Both silages in Exp1 and Exp2 had relatively low CP contents. In Exp1, low proportion of
clover was not able to compensate for the shortage of nitrogen fertilization. When targeting to
optimize raw material species, fertilization and growth stage should be considered. Further,
proteolysis during ensiling should be balanced so that enough solubilisation takes place while
breakdown of amino acids is restricted. Highest proportion of protein was separated when
using no additive, but this resulted in protein having less lysine than in the additive treated
silage juice. This is in line with the results of Winters et al. (2001), who suggested that
additives aid to achieve better amino acid profile in silage than without additives.
Potential to reduce phosphorus load
High amount of phosphorus is silage effluent has given it a bad reputation in relation to
environmental load. When targeting to produce silage for pigs, high solubility of phosphorus
in silage could be exploited to reduce nutrient runoff to the waterways. The effect would be
achieved in three ways:
1) In Finland, typically more phosphorus is removed from soil by harvested grass silage than
by harvested grain yield.
2) Nowadays, feeds typically bring more phosphorus to the farm than will be sold as animal
products. On farm cultivated grass extract could replace part of the monocalcium phosphate
and soybean meal in pig diets, thus reducing phosphorus accumulation.
3) Grass cultivation is encouraged (by subsidies), especially on riversides and lakesides, to
avoid erosion. If there is no use for the grass and it is not harvested, phosphorus within the
biomass is in great danger to leak into the water. An efficient usage for grass biomass would
alleviate that problem.
Economic feasibility
In Exp2, the value of concentrated silage extract was calculated to be 381 €/t DM, when
included in diets (10 % DM) to fattening pigs (Table 3). To achieve economic feasibility, the
gain of the separation process should be improved (e.g. from 25 to 40 % for DM and from 20
to 40 % for CP) while keeping the process simple. Further, the economic value of reduced
phosphorus load should be included into the calculations. While the calculated price of silage
extract already covers the production cost of grass silage for biogas production (80 €/t DM,
not including the silos (Kässi and Aro-Heinilä, 2014), economic feasibility might be
achievable in near future.
166
Proceedings of the 5th Nordic Feed Science Conference
Methods and miscellaneous
The fibrous residue of processing could be used e.g. as stimulus feed for pregnant sows, high
quality feed for horses or as a source of energy. For horse owners, a feed with positive health
effects could be marketed at a higher price. The reduction in protein and sugar as a
consequence of fractioning would be beneficial for animals with low energy requirements.
Table 3 Recipes for liquid feeding with and without silage extract.
Two recipes for liquid pig feeding (% of feed DM)
Feed components
Price, €/t
Without silage extract
Silage extract included
(10 % DM)
Barley
140
73
65
Wheat
165
10
10
13.3
11
Soybean meal (GMO-free) 500
Silage extract
204
Rapeseed oil
1250
0.57
1.5
L-Lysin
1400
0.34
0.37
L-Threonin
2000
0.08
0.08
DL-Methionine
2750
0.03
0.07
Monocalcium phosphate
500
1.04
0.78
Calcium carbonate
75
1.59
1.47
Added acid
2000
Dosage 2 l/liquid feed
-
211
211
Price of the mixture, €/t
10
Conclusions
Processing grass silage for pig feeding could improve protein self sufficiency and reduce
phosphorus leakage into watercourses. Economic sufficiency could be achieved by
combining all positive effects. More research is however needed on several aspects of the
process and further on the effects on pig production and environment.
References
Choi, C.-W., Vanhatalo, A. & Huhtanen, P., 2003. Effects of type of grass silage and level of
concentrate on the flow of soluble non-ammonia nitrogen entering the omasum of dairy cows.
J. of Anim. feed Sci. 12, 3-22.
Ecker, J., Schaffenberger, M, Koschuh, W., Mandl, M., Böchzelt, H.B., Schnitzer, H.,
Harasek, M. & Steinmüller, H., 2012. Green Biorefinery Upper Austria – Pilot Plant
operation. Sep. Purif. Technol. 96, 237–247.
Grass, S., 2004. Utilisation of grass for production of fibres, Protein and energy. In: Biomass
and Agriculture: Sustainability, Markets and Policies. OECD Publication services, Paris.
Available at: http://www.bpsag.ch/pdf/Publication%20OECD.pdf. (Accessed May 1, 2014).
Huhtanen, P., Nousiainen, J. & Rinne, M., 2006. Recent developments in forage evaluation
with special reference to practical applications. Agric. Food Sci. 15, 293-323.
Kamm, B. & Kamm, M., 2004. Principles of biorefineries. Appl. Microbiol. Biotechnol. 64,
137–145.
Proceedings of the 5th Nordic Feed Science Conference
167
Methods and miscellaneous
Kammes, K.L., Bals, B.D., Daleb,B.E. & Allena, M.S., 2011. Grass leaf protein, a coproduct
of cellulosic ethanol production, as a source of protein for livestock. Anim. Feed Sci.
Technol. 164, 79–88.
Keijsers, E. & Mandl, M., 2010. Green Biorefinery, IEA Biorefinery Course. Available at:
http://www.bioref-integ.eu/fileadmin/bioref-integ/user/documents/8._KeijsersGreen_Biorefinery_130910.pdf (cited May 1, 2014).
Kässi, P., Aro-Heinilä, E. 2014. Nurmibiokaasulaitoksen talous. In: Seppälä, A., Kässi, P.,
Aro-Heinilä, E., & Lehtonen, H. 2014. Nurmesta biokaasua liikennepolttoaineeksi.
Bionurmi-hankkeen loppuraportti. Available at www.mtt.fi/bionurmi.
Mandl, M.G., 2010. Status of green biorefining in Europe. Biofuels, Bioproducts and
Biorefining 4, 268–274.
McDonald, P., Henderson A.R. & Heron S.J.E., 1991. The Biochemistry of Silage. 2nd
edition, Chalcombe Publications, Marlow, UK. 340 p.
O’Keeffe, S., Schulte R.P.O., Sanders, J.P.M. & Struik, P.C., 2011. I. Technical assessment
for first generation green biorefinery (GBR) using mass and energy balances: Scenarios for
an Irish GBR blueprint. Biomass and bioenergy 35, 4712 – 4723.
Patterson, DC. & Kilpatrick, D.J., 1991. Effluent from grass-silage for finishing pigs. J.
Agric. Sci., 116, 119-124.
Presto, M., Rundgren, M. & Wallenbeck, A. 2013. Inclusion of grass/clover silage in the diet
of growing/finishing pigs – Influence on pig time budgets and social behaviour. Acta Agric.
Scand., Section A - Anim. Sci. 63, 84-92
Seppälä, A., Kässi, P., Aro-Heinilä, E. & Lehtonen, H. 2014. Nurmesta biokaasua
liikennepolttoaineeksi. Bionurmi-hankkeen loppuraportti. Available at www.mtt.fi/bionurmi.
Sieker, T., Neuner, A., Dimitrova, D., Tippkötter, N., Muffler, K., Bart, H.-J., Heinzle, E. &
Ulber, R., 2011. Ethanol production from grass silage by simultaneous pretreatment,
saccharification and fermentation: First steps in the process development. Engineering in Life
Sciences 11, 436-442.
Winters, A.L., Fychan, R. & Jones, R., 2001. Effect of formic acid and a bacterial
inoculanton the amino acid composition of grass silage and on animal performance. Grass
Forage Sci. 56, 181–192.
168
Proceedings of the 5th Nordic Feed Science Conference
Methods and miscellaneous
Dry matter yields and feed values of faba bean and pea as bi-crops with wheat or oats in
Northern Finland
K. Kuoppala1, A. Huuskonen2, E. Saarinen3 & M. Rinne1
1
MTT Agrifood Research Finland, Animal Production Research, Animale, FI-31600
Jokioinen, Finland; 2MTT Agrifood Research Finland, Animal Production Research,
Halolantie 31 A, FI-71750 Maaninka; 3MTT Agrifood Research Finland, Animal Production
Research, Tutkimusasemantie 15, FI-92400 Ruukki
Correspondence: [email protected]
Introduction
Producing whole crop small grain cereal silages (WCS) provides an opportunity to improve
the efficiency of forage production for ruminants under Nordic conditions. Harvesting and
storage of WCS is possible with the same methods as used for grass silage so that only one
machinery chain is needed on the farm. Using on-farm grown cereals as silages rather than as
grains results in significant economic benefits, when combine harvesting and drying of the
grains are not needed (Turunen, 2003). The nutritional characteristics of WCS are similar to
those of maize silage, and the Nordic growing conditions are much better suited for it.
The WCS can be produced from pure stands of typically wheat or barley, but inclusion of
grain legumes (faba bean, pea) or vetches is of interest particularly in organic production.
Inclusion of legumes allows reductions in nitrogen fertilization, increase biodiversity and
improve soil structure.
Inclusion of legumes into WSC may also improve the nutritional quality of the feed. The CP
concentration of cereal based WCS is typically around 100 g/kg DM or even lower
(Huuskonen & Joki-Tokola, 2010; MTT 2014). Finnish protein feeding recommendations for
growing cattle (MTT, 2014) are not usually fulfilled if whole-crop silage based rations are
fed without protein supplementations. By including legumes into the mixture, the CP
concentrations may rise to approximately 130-150 g/kg DM. Such concentration is already
high enough to be used for growing cattle without separate protein supplementation
(Huuskonen et al., 2007; Huuskonen, 2009). According to Finnish feeding recommendations
the protein intake for growing cattle over 200 kg live weight is considered adequate if the
protein balance in the rumen (PBV) of the total diet is not lower than -10 g/kg DM (MTT,
2014).
In this paper the composition, digestibility and dry matter (DM) yield of legume and cereal
bi-crops are discussed in order to better understand the basis of the feed value of whole crop
silages made of legumes and cereals grown together, and their relevance to cattle feeding
under Nordic conditions.
Materials and Methods
Three cultivars of faba bean (Vicia faba) and four cultivars of pea (Pisum sativum) were
cultivated in a plot experiment at MTT Ruukki (64°N) in Northern Finland. Both legumes
were cultivated as bi-crops with spring wheat (Triticum aestivum cv. Wappu) and with oats
(Avena sativa cv. Wilhelmiina) as four replicates. The cultivars of faba bean were Kontu,
Fuego and Tangenta and of pea Arvika, Dolores, Florida and Jermu. The sowing dates were
29 and 30 May for faba bean and pea, respectively. Seeding rates were: faba bean 170, pea
156, wheat 80 and oats 70 kg/ha. The bi-crop plots were harvested with a plot harvester
(Haldrup) three times: 16 August, 30 August and 13 September in 2012 for faba bean and 15
Proceedings of the 5th Nordic Feed Science Conference
169
Methods and miscellaneous
August, 27 August and 10 September for pea. The yield was measured by weighing the
whole yield harvested from each plot.
Samples of one replicate of each bi-crop and also from the samples after manual separation of
different plant species were analysed in the laboratory of MTT. The DM concentration was
determined by drying at 105 °C for 20 h and of organic matter (OM) by ashing at 600 °C for
2 h. Crude protein (CP) concentration was analysed by the Dumas method using Leco FP 428
nitrogen analyser (Leco Corp., St Joseph; USA). Starch was analysed according to Salo &
Salmi (1968). The concentration of neutral detergent fibre (NDF) was determined according
to Van Soest et al. (1991) using Na sulphite, without amylase and with ash excluded. The
concentration of indigestible NDF (iNDF) was determined by a 12 d ruminal incubation in
the rumen of dairy cows fed a forage-based diet and using nylon bags with a pore size of 17
μm (Huhtanen et al., 1994) and expressed ash-free.
Digestibility of OM (OMD) was determined using in vitro pepsin-cellulase method and
general correction equation was used to convert pepsin-cellulase solubility values into in vivo
digestibility according Huhtanen et al. (2006). D-value (the concentration of digestible OM,
g/kg DM) was calculated using the analysed ash concentration. The results are presented for
both legume bi-crops as means of the cultivars and the two cereal grain species and for pure
plant species separately. Statistical analyses were conducted with ANOVA with legume bicrop (pea or faba bean) and harvesting date in the model (Table 1) and with plant species
(pea, faba bean, oats and wheat) and harvesting date in the model (Table 2). Pairwise
comparisons were made with Tukey t-test.
Results and Discussion
Harvesting later increased the DM concentration of the bi-crops of both legumes (Table 1).
The DM yield was higher for faba bean than for pea bi-crops and it increased for both legume
cereal bi-crops when harvested later. Highest DM yield was recorded for pea cv. Dolores and
faba bean cv. Fuego (data not shown). The average proportion of legume in the bi-crop was
higher for pea (mean 0.62) than for faba bean (mean 0.49).
The CP concentration was higher for pea bi-crops than for faba bean which may originate
from the higher proportion of pea in the bi-crop compared with faba bean. The CP
concentration of pea bi-crops did not change significantly with postponed harvest but that of
faba bean increased from first to second harvest. When analysed separately the pure legume
whole crops had the average CP concentrations of 167 and 163 g/kg DM for pea and faba
bean, respectively (Table 2). The CP concentration increased for faba bean but did not change
for pea whole crops. The average CP concentration of pure whole crop cereals was low, 109
and 106 g/kg DM for oats and wheat, respectively. Crude protein yield was on the average
1194 and 1274 kg/ha for pea and faba bean bi-crops, respectively and it increased for both
legumes when the harvest was postponed.
170
Proceedings of the 5th Nordic Feed Science Conference
Methods and miscellaneous
Table 1 The effect of harvesting date for performance and chemical composition of pea cereal and faba bean
cereal bi-crops
Pea cereal bi-crop (n=8)
Harvest date in 2012
DM concentration
1
189
2
a
DM yield, kg/ha
7193
Proportion of legume
0.54
Crude protein yield, kg/ha
1053
187
a
3
a
7437
239
ab
0.65
a
Faba bean cereal bi-crop (n=6)
1173
SEM
b
8867
6.4
b
0.68
ab
1356
b
1
182
2
a
254.1
7717
0.032
0.46
206
a
a
37.5
992
2.1
60.1a
3
ab
9713
b
0.50
1402
218
SEM
b
9881
4.2
b
0.50
ab
1429
262.4
0.016
b
61.7
Chemical composition, g/kg DM
Ash
73.6a
a
72.6ab
160
a
Crude protein
147
Starch
31.6a
62.1ab
NDF
503a
458b
a
a
185
65.1b
154
a
a
59.7a
1.41
128
94.8b
11.17
86.1a
105.1a 151.1b 6.2
471ab
11.9
499a
475ab
184
170
472b
7.9
b
4.1
188
iNDF/NDF
0.376a
0.404a 0.390a 0.0109
0.400a 0.358b 0.364b 0.0078
OMD, g/g
0.662a
0.690b 0.699b 0.0074
0.657a 0.691b 0.696b 0.0055
D-value, g/kg DM
613a
640b
618a
7.7
200
b
4.3
iNDF
653c
4.7
a
144
b
7.2
a
144
b
57.1a
650b
172
656b
5.6
Harvesting dates for pea bi-crops: 1 =August, 15; 2 = August, 27; 3 = September, 3; and for faba bean bi-crops:
1 = August, 16; 2 = August, 30; 3 = September, 13. DM=dry matter, NDF=neutral detergent fibre,
iNDF=indigestible NDF, OMD= digestibility of organic matter, D-value=concentration of digestible organic
matter in dry matter. Same superscript on the same row within the plant means no statistical significance
between harvesting dates, P>0.05.
The starch concentration was lower for pea than for faba bean bi-crops and it increased with
postponed harvest for both. According to Salawu et al. (2001) the increase in starch
concentration means that the growth is changing from vegetative phase to generative phase
and preparing the seeds.
The NDF and iNDF concentrations were on average 477 and 482 g/kg DM and 185 and 181
g/kg DM for pea and faba bean bi-crops, respectively. The NDF concentrations of these
legume bi-crops were on the same level or lower than for early cut grass herbages but higher
than for red clover whereas the concentration of iNDF was much higher than in grass or red
clover (Kuoppala et al., 2009). Red clover typically has lower NDF and higher iNDF
concentration compared to grasses both in primary growth and regrowth (see Kuoppala, 2010
for review). In that review the average proportion of totally indigestible part in NDF, i.e.
iNDF, was calculated to be 0.35 for red clover and only 0.18 for grasses. In the present data it
was higher, 0.39 for pea bi-crops and 0.37 for faba bean bi-crops. For pea bi-crops it did not
change whereas for faba bean bi-crops it decreased when harvest was postponed. When the
plant species were analysed separately (Table 2) the concentration of NDF was higher for the
cereals than for the legumes whereas that of iNDF was slightly higher or at the same level.
High proportion of iNDF in NDF seems to be common to the grain legumes pea and faba
bean as well as forage legumes (Kuoppala, 2010).
Proceedings of the 5th Nordic Feed Science Conference
171
Methods and miscellaneous
Table 2 The effect of harvesting time for chemical composition (g/kg DM) of pea, faba bean, oats and wheat separated from the whole bi-crop
samples.
Plant
Pea (n=4)
Harvest
1
Ash
CP
61.9
172
2
61.6
165
Faba bean (n=3)
3
56.4
164
SEM
1
2
a
3.66
55.7
11.1
a
147
3
50.4
159
Oats (n=2)
b
ab
55.5
182
a
b
Wheat (n=2)
SEM
1
2
3
SEM
1
2
3
SEM
0.96
81.2
86.6
81.6
8.62
69.4
69.0
75.8
4.98
8.0
115
112
100
7.8
103
107
a
86.2
107
b
118.3
7.6
b
Starch
40.1
73.5
86.8
13.5
60.4
94.3
90.4
9.1
46.4
86.5
124.9
20.82
22.3
8.07
NDF
403
398
411
15.9
420
425
398
8.39
559
562
541
14.7
563
551
584
11.9
iNDF
184
163
162
10.7
181
165
146
11.9
190
189
186
3.22
187
190
189
1.8
iNDF/NDF 0.455 0.409 0.393 0.0165
0.430 0.438 0.367 0.0191
0.336 0.344 0.332
0.023
0.332 0.345
0.324
0.023
OMD, g/g
0.706 0.730 0.732 0.0128
0.704 0.722 0.740 0.0107
0.639 0.639 0.635
0.0491
0.623 0.635
0.628
0.039
D value
662
665
587
9.31
579
581
4.7
685
691
13.4
686
699
9.74
584
584
591
Harvesting dates for pea bi-crops: 1 =August, 15; 2 = August, 27; 3 = September, 3; and for faba bean bi-crops: 1 = August, 16; 2 = August, 30; 3 =
September, 13. DM=dry matter, CP=crude protein, NDF=neutral detergent fibre, iNDF=indigestible NDF, OMD= digestibility of organic matter, Dvalue=concentration of digestible organic matter in dry matter. Same superscript on the same row within the plant means no statistical significance
between harvesting dates, P>0.05.
172
Proceedings of the 5th Nordic Feed Science Conference
Methods and miscellaneous
The digestibility expressed as OMD or D-value was rather low for both legume bi-crops
(Table 1). Opposite to grasses it increased with the postponed harvest. The D-values
corresponded to the values of very late cut grass but were on the same level as the average
values for red clover (Kuoppala, 2010). When the plant species were analysed separately
(Table 2) it seems that the pure legume whole crops’ digestibility was higher than that of the
cereals, being on the same level as good quality grass silage at least in the two later harvest
dates.
Inclusion of forage legumes or WCS into the diet of dairy cows has been shown to increase
the voluntary feed intake of dairy cows (Huhtanen et al., 2007). Cows eat more silage when
red clover is mixed with grasses in the herbage in spite of the higher concentration of iNDF.
It could be assumed that the situation will be same for pea and faba bean also. There are also
positive experiences of using ensiled cereal + legume mixtures as feeds for dairy cows (Haag,
2007; Juutinen, 2011).
It is suggested that the starch concentration of whole-crop silages, together with high starch
digestibility could compensate for the high concentration of indigestible fibre (Wallsten,
2008). Both dairy cows (Ahvenjärvi et al., 2006) and finishing beef cattle (Keady et al.,
2013) have often been able to maintain or even increase silage dry matter intake (DMI) after
inclusion of whole-crop silage into the diet, although the digestibility of whole-crop silage
has been lower than that of the grass silage. According to Huhtanen et al. (2007), the dairy
cows responses to replacing grass silages partially or totally with whole-crop silages on DMI
could not be accurately predicted from differences in silage D-value, DM concentration or
fermentation characteristics. The maximum silage DMI increase was obtained when the
proportion of the whole-crop silage was 0.48 of total silage DM, and the effect was quadratic
(Huhtanen et al., 2007).
Conclusions
Postponing the harvest increased dry matter and crude protein yields and digestibility of the
faba bean cereal and pea cereal bi-crops. Inclusion of legumes has a potential to increase
energy and protein values of whole crop cereal silages, because they were higher in the
legume fractions compared to the cereals.
References
Ahvenjärvi, S., Joki-Tokola, E., Vanhatalo, A., Jaakkola, S. & Huhtanen, P., 2006. Effects of
replacing grass silage with barley silage in dairy cow diets. J Dairy Sci 89, 1678–1687.
Haag, T., 2007. Åkerböna i samodling med vårvete som helgrödesensilage till mjölkkor.
[Field beans cultivated together with spring wheat as whole-crop silage to dairy cows]. SLU
Institutionen för norrländsk jordbruksvetenskap. Rapport 3, 2007 (In Swedish).
Huhtanen, P., Kaustell, K. & Jaakkola, S., 1994. The use of internal markers to predict total
digestibility and duodenal flow of nutrients in cattle given six different diets. Anim. Feed Sci.
Technol. 48, 211–227.
Huhtanen, P., Rinne, M. & Nousiainen, J., 2007. Evaluation of the factors affecting silage
intake of dairy cows: a revision of the relative silage dry-matter intake index. Animal 1, 758770.
Proceedings of the 5th Nordic Feed Science Conference
173
Methods and miscellaneous
Huuskonen, A., 2009. The effect of cereal type (barley versus oats) and rapeseed meal
supplementation on the performance of growing and finishing dairy bulls offered grass
silage-based diets. Livest. Sci. 122, 53–62.
Huuskonen, A. & Joki-Tokola, E., 2010. Performance of growing dairy bulls offered diets
based on silages made of whole-crop barley, whole-crop wheat, hairy vetch and grass. Agric.
Food Sci. 9, 116-126.
Huuskonen, A., Khalili, H. & Joki-Tokola, E., 2007. Effects of three different concentrate
proportions and rapeseed meal supplement to grass silage on animal performance of dairybreed bulls with TMR feeding. Livest. Sci. 110, 154-165.
Juutinen, E. 2011. Säilörehua herneestä ja härkäpavusta. Nauta 4/11, pp. 34-35. (In Finnish).
Keady, T.W.J., Hanrahan, J.P., Marley, C.L. & Scollan, N.D., 2013. Production and
utilization of ensiled forages by beef cattle, dairy cows, pregnant ewes and finishing lambs: a
review. Agric. Food Sci 22, 70–92.
Kuoppala, K. 2010. Influence of harvesting strategy on nutrient supply and production of
dairy cows consuming diets based on grass and red clover silage. MTT Science 11: 50 s.
Doctoral Dissertation. Available at: www.mtt.fi/mtttiede/pdf/mtttiede11.pdf
Kuoppala, K., Ahvenjärvi, S., Rinne, M. & Vanhatalo, A., 2009. Effects of feeding grass or
red clover silage cut at two maturity stages in dairy cows. 2. Dry matter intake and cell wall
digestion kinetics. J. Dairy Sci. 92, 5634-5644.
MTT. 2013. Finnish Feed Tables and Feeding Recommendations – web service. Available at:
www.mtt.fi/rehutaulukot/english.
Salawu, M.B., Adesogan, A.T., Weston & C.N., Williams, S.P., 2001. Dry matter yield and
nutritive value of pea-wheat bi-crops differing in maturity at harvest, pea to wheat ratio and
pea variety. Anim. Feed Sci Technol. 94, 77-87.
Turunen, H., 2003. Kokoviljasäilörehun taloudellisuus nautakarjatilalla. In: Lampinen, K.,
Harmoinen, T. & Teräväinen H. (Eds.). Kokoviljasäilörehun tuotanto ja käyttö. Jyväskylä,
Finland: Gummerus Kirjapaino Oy. pp. 5-16. (In Finnish).
Wallsten, J., 2008. Whole-crop cereals in dairy production. Digestibility, feed intake and milk
production. Doctoral Thesis. Swedish University of Agricultural Sciences, Umeå, Sweden. 45
p.
174
Proceedings of the 5th Nordic Feed Science Conference
DISTRIBUTION:
Sveriges Lantbruksuniversitet
Institutionen för husdjurens utfodring och vård
Box 7024
750 07 UPPSALA
Tel.018/672817
[email protected]